Solution Deploy Network Solution optional

This guide explains how to install the WeaveNet pod networking solution on your Kubernetes cluster and troubleshoot why an application pod might be stuck in the "ContainerCreating" state. We discuss inspecting pod status, deploying the network solution, customizing configuration settings, and verifying that everything is operating correctly.

Inspecting the Application Pod

An application named "app" has been deployed in the default namespace. Begin by checking the pod status:

kubectl get pods

If you see that the pod is stuck in the "ContainerCreating" state, it indicates that the pod is not running. To diagnose further, describe the pod to review its events:

kubectl describe pod app

Focus on the events section at the bottom of the output. You may encounter an error message like:

Warning  FailedCreatePodSandbox  35s   kubelet  Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox "d25e340bdc7fb268f261540c1bcd937e356ccae1fa9f2ef621e3f7e89": plugin type="weave-net" name="weave" failed (add): unable to allocate IP address: Post "http://127.0.0.1:6784/ip/add?…": dial tcp 127.0.0.1:6784: connect: connection refused

This error typically means that the network is not configured properly because the container runtime was unable to allocate an IP address.

Deploying WeaveNet

To fix the network configuration issue, deploy the WeaveNet networking plugin to your cluster. For more background on installing add-ons, see the Kubernetes documentation on installing add-ons.

Follow these steps:

1. Open the Kubernetes add-ons documentation, which covers various networking add-ons like ACI, Antrea, and Calico.

   

2. Install WeaveNet by applying the manifest file with this command:

   kubectl apply -f https://github.com/weaveworks/weave/releases/download/2.8.1/weave-daemonset-k8s.yaml
   

3. Cluster CIDR Considerations

   Review your cluster's CIDR settings. If you have set a specific cluster CIDR in your kube-proxy (for example, clusterCIDR: 10.244.0.0/16), ensure that WeaveNet is configured to use a compatible IP allocation range. List the WeaveNet pods and check their logs to verify:

   kubectl get pod -n kube-system | grep weave
   kubectl logs <weave-net-pod> -n kube-system
   

If the default manifest does not fit your cluster settings, adjust the configuration. For instance, to set the IP allocation range to "10.0.0.0/24", update the environment variables in the manifest’s container configuration:

containers:
  - name: weave
    env:
      - name: IPALLOC_RANGE
        value: "10.0.0.0/24"

Verifying and Customizing the WeaveNet Manifest

To further customize WeaveNet, first download the manifest:

wget https://github.com/weaveworks/weave/releases/download/v2.8.1/weave-daemonset-k8s.yaml

Open the file to review the container definitions. You will find a section like:

metadata:
  labels:
    name: weave-net
spec:
  initContainers:
    - name: weave-init
      image: 'weaveworks/weave-kube:latest'
      imagePullPolicy: Always
      command:
        - /home/weave/init.sh
      securityContext:
        privileged: true
      volumeMounts:
        - name: cni.bin
          mountPath: /host/opt
        - name: cni-bin2
          mountPath: /host/home
        - name: cni/conf
          mountPath: /host/etc
        - name: lib/modules
          mountPath: /lib/modules
        - name: xtables-lock
          mountPath: /run/xtables.lock
    - name: weave
      image: 'weaveworks/weave-kube:latest'
      imagePullPolicy: Always
      command:
        - /home/weave/launch.sh
      env:
        - name: INIT_CONTAINER
          value: "true"
        - name: HOSTNAME
          valueFrom:
            fieldRef:
              apiVersion: v1
              fieldPath: spec.nodeName

To specify the IP allocation range for WeaveNet, add the environment variable under the container’s env section:

      env:
        - name: INIT_CONTAINER
          value: "true"
        - name: HOSTNAME
          valueFrom:
            fieldRef:
              fieldPath: spec.nodeName
        - name: IPALLOC_RANGE
          value: "10.244.0.0/16"

After making the necessary modifications, apply the updated manifest:

kubectl apply -f weave-daemonset-k8s.yaml

Verifying the Deployment

After applying the manifest, ensure that the WeaveNet pods are running successfully in the kube-system namespace:

kubectl get pods -n kube-system

You should see an output similar to:

NAME             READY   STATUS    RESTARTS   AGE
weave-net-xxxxx  2/2     Running   0          10s

This output confirms that WeaveNet has been deployed successfully as a DaemonSet across your nodes.

Additional Diagnostics

If you encounter issues such as evicted pods or ongoing network problems, use the following commands to gather more information:

kubectl get evicted
kubectl get pods

Review the events for issues like resource constraints or network connectivity problems. Additionally, inspect the kube-proxy configuration with:

kubectl describe configmap kube-proxy -n kube-system

Ensure that the configuration includes a matching cluster CIDR, such as:

apiVersion: kubeproxy.config.k8s.io/v1alpha1
...
clusterCIDR: 10.244.0.0/16

Make sure any modifications in the WeaveNet manifest align with your cluster’s CIDR settings.

Conclusion

Once WeaveNet is deployed and confirmed to be running, recheck the status of the "app" pod:

kubectl get pods

If the network configuration is correct, the pod should transition from the "ContainerCreating" state to Running. This completes the exercise on deploying the WeaveNet networking solution and troubleshooting pod creation issues.