Skip to main content

Troubleshooting

kube-flannel CrashLoopBackOff​

Symptom​

Flannel pods crash with error:

Failed to check br_netfilter: stat /proc/sys/net/bridge/bridge-nf-call-iptables: no such file or directory

Solution​

Load the kernel module on the affected node:

modprobe br_netfilter

Then restart flannel pods:

kubectl delete pod -n kube-flannel <flannel-pod-name> --force

Prevention​

Add modprobe br_netfilter to your node provisioning steps.

MetalLB CRDs OutOfSync​

Symptom​

ArgoCD shows MetalLB CRDs as OutOfSync.

Cause​

MetalLB v0.14+ uses CRDs with webhook conversions. The webhook CA bundle is dynamically generated.

Solution​

The metallb.yaml template includes ignoreDifferences to ignore the caBundle field:

ignoreDifferences:
  - group: apiextensions.k8s.io
    kind: CustomResourceDefinition
    name: bgppeers.metallb.io
    jsonPointers:
      - /spec/conversion/webhook/clientConfig/caBundle

This is already configured in the template.

ArgoCD Repo-Server OOM​

Symptom​

ArgoCD repo-server pods crash with OOM errors.

Solution​

Configure resources in the ArgoCD values:

repoServer:
  resources:
    requests:
      cpu: 250m
      memory: 256Mi
    limits:
      cpu: 1000m
      memory: 1Gi

Node Not Ready​

Common Causes​

  1. Kubelet not running:

    systemctl status kubelet
    systemctl restart kubelet

  2. Certificates missing: Re-run kubeadm join on the node

  3. Network issues: Check that br_netfilter module is loaded and sysctl settings are applied

Deploy Not Working​

Check 1: ArgoCD Applications​

kubectl get applications -n argocd

Check 2: Application Revision​

kubectl get application <app-name> -n argocd -o jsonpath='{.spec.source.targetRevision}'

Check 3: Sync Status​

kubectl get application <app-name> -n argocd -o jsonpath='{.status.sync.status}'

Check 4: ArgoCD Logs​

kubectl logs -n argocd deployment/argocd-application-controller