Fix Minikube When It Breaks - A 3AM Debugging Guide

This happens way too often.

Usually it's one of these issues:First thing to try is the nuclear option:bashminikube deleteminikube start --driver=docker --memory=4096Works half the time, saves you an hour of debugging.

Still fucked? Check if virtualization is enabled in your BIOS:```bash# Linux

• should return a number > 0egrep -c '(vmx|svm)' /proc/cpuinfo# macOS
• should show VMX supportsysctl -a | grep machdep.cpu.features```Still stuck?

Try different drivers:bashminikube start --driver=virtualbox # Slow but reliableminikube start --driver=hyperv # Windows onlySometimes the first driver just won't work, even if it worked yesterday. Driver bullshit is common.

Time to wrestle with your BIOS.

Reboot and hit F2, F12, Delete, or whatever key gets you into BIOS setup (it's different for every manufacturer, naturally).Look for virtualization settings:

• Intel machines: "Intel VT-x", "Virtualization Technology", or "VMX"
• AMD machines: "AMD-V", "SVM Mode", or similar

Enable it, save changes, reboot.

Some corporate laptops have this locked down

• if you can't change BIOS settings, try the Docker driver instead:bashminikube start --driver=dockerThe Docker driver doesn't need hardware virtualization, though it has its own set of quirks.

Usually means your driver is broken:bash# Check what drivers you haveminikube config get driver# Try Docker (most reliable)minikube deleteminikube start --driver=docker# If Docker fails, try VirtualBoxminikube start --driver=virtualbox

HyperKit on macOS is flaky as hell:bash# Switch to VirtualBox (slower but stable)brew install virtualboxminikube deleteminikube start --driver=virtualbox# Or use Docker (fastest)minikube start --driver=docker

Two common issues:

1. Docker Desktop conflict
   • Stop Docker Desktop before starting Minikube
2. Windows features
   • Ensure "Hyper-V" AND "Windows Hypervisor Platform" are enabledpowershell# Run as AdminEnable-WindowsOptionalFeature -Online -FeatureName Microsoft-Hyper-V -All

This is usually one of three problems:

Private registry without credentials:

## Check the exact error
kubectl describe pod <podname>

## If it mentions authentication, create registry secret
kubectl create secret docker-registry regcred \
  --docker-server=your-registry.com \
  --docker-username=your-user \
  --docker-password=your-pass

## Add to your pod spec
imagePullSecrets:
- name: regcred

Image doesn't exist:

## Test if image exists
docker pull your-image:tag

## Common mistakes:
## - Wrong tag (using "latest" but pushed "v1.0")
## - Typo in image name
## - Image exists in different registry

Local images not available in cluster:

## Use Minikube's Docker daemon
eval $(minikube docker-env)
docker build -t myapp:latest .

## Or load image into cluster
minikube image load myapp:latest

Your pods are running out of memory. Check what's actually happening:

## See resource usage
kubectl top pods

## Check limits in your deployment
kubectl describe deployment myapp

## See events
kubectl get events --sort-by=.metadata.creationTimestamp

Quick fixes:

## Increase pod memory limit
kubectl patch deployment myapp -p '{"spec":{"template":{"spec":{"containers":[{"name":"myapp","resources":{"limits":{"memory":"512Mi"}}}]}}}}'

## Or increase Minikube memory
minikube delete
minikube start --memory=8192

Minikube's VM is full. Found out during a demo when kubectl apply started throwing no space left on device in front of everyone. Usually it's:

• Docker images from all the testing you've been doing
• Application logs that just keep growing
• PersistentVolumes you forgot you created

Clean house:

## SSH into Minikube
minikube ssh

## Check disk usage
df -h

## Clean Docker images
docker system prune -a -f

## Check large files
sudo find / -type f -size +100M 2>/dev/null

Permanent fix:

## Restart with more disk space
minikube delete
minikube start --disk-size=50GB

Linux file descriptor limits are too low:

## Check current limits
ulimit -n

## Increase temporarily
ulimit -n 4096

## Permanent fix - add to ~/.bashrc
echo "ulimit -n 4096" >> ~/.bashrc

Scheduler can't find a node with enough resources:

## Check node resources
kubectl describe node minikube

## Look for resource pressure
kubectl get nodes -o wide

## Check pod requirements
kubectl describe pod <podname>

Common causes:

• Requesting more memory than node has
• Node has resource pressure (disk/memory)
• Affinity rules blocking scheduling

Quick fix:

## Reduce resource requests
kubectl patch deployment myapp -p '{"spec":{"template":{"spec":{"containers":[{"name":"myapp","resources":{"requests":{"memory":"128Mi","cpu":"100m"}}}]}}}}'