Microk8s启用DNS后Pod卡在ContainerCreating状态求助

Looking at the logs you shared, the core issue is obvious: your kubelet can’t sync the secret and ConfigMap caches, which leads to volume mounting timeouts for the CoreDNS pod. Let’s walk through practical, actionable steps to diagnose and fix this:

Step 1: Check Health of Core MicroK8s Components #

First, make sure the foundational services are running properly:

• Verify the kubelet status:

  systemctl status snap.microk8s.daemon-kubelet
  

• Check containerd (the container runtime MicroK8s uses):

  systemctl status snap.microk8s.daemon-containerd
  

If either service is stopped or throwing errors, restart them with:

sudo systemctl restart snap.microk8s.daemon-kubelet
sudo systemctl restart snap.microk8s.daemon-containerd

Step 2: Test API Server Connectivity #

The kubelet relies on the Kubernetes API Server to fetch secrets and ConfigMaps. Test if your node can reach the API Server:

• Use the cluster IP from your kubectl get services output (10.152.183.1):

  curl -k https://10.152.183.1:443
  

A successful response will return JSON listing API versions. If this fails, check the API Server logs for issues:

microk8s.kubectl logs -n kube-system pod/kube-apiserver-peza-ubuntu-19

Step 3: Clear Corrupted Kubelet Cache #

Corrupted cache is a frequent cause of volume mounting failures. Let’s clean it up:

1. Stop MicroK8s:

   microk8s stop
   

2. Delete the kubelet cache directory:

   sudo rm -rf /var/snap/microk8s/current/var/lib/kubelet/cache
   

3. Restart MicroK8s:

   microk8s start
   

Wait 1-2 minutes, then check if CoreDNS starts:

microk8s.kubectl get pods -n kube-system

Step 4: Recreate CoreDNS Resources #

If the issue persists, the CoreDNS service account or associated secrets/ConfigMaps might be corrupted. Reset the DNS add-on:

1. Disable DNS:

   microk8s.disable dns
   

2. Delete the CoreDNS service account (forces a new token secret to be created):

   microk8s.kubectl delete serviceaccount coredns -n kube-system
   

3. Re-enable DNS:

   microk8s.enable dns
   

Step 5: Troubleshoot VirtualBox Network #

Since you’re running Ubuntu in VirtualBox, network glitches between the node and API Server could be blocking communication:

• Confirm the node’s IP (10.0.2.15) is reachable and not conflicting with other devices.
• Temporarily disable Ubuntu’s firewall to rule out blocking:

  sudo ufw disable
  

• Check if iptables rules are interfering with Kubernetes traffic:

  iptables -L -n | grep -E "kube|coredns"
  

If you see unexpected block rules, reset iptables (note: this clears all custom rules):

sudo iptables -F && sudo iptables -t nat -F && sudo iptables -t mangle -F && sudo iptables -X

After trying these steps, your CoreDNS pod should transition to a Running state. If it still doesn’t work, share the updated kubectl describe pod output and we can dig deeper.

内容的提问来源于stack exchange，提问作者Peter Zandbergen