Files
myceliumcloud-examples/examples/nginx-nodeport/deployment-troubleshooting.md

355 lines
9.8 KiB
Markdown

# Deployment Troubleshooting Guide for nginx-nodeport
This guide helps resolve common connectivity and deployment issues with the nginx-nodeport implementation.
## 🚨 Common Connection Issues
### Issue 1: API Server Timeout
**Error**: `dial tcp [IPv6]:6443: i/o timeout`
**Cause**: Kubernetes API server is slow to respond or network connectivity issues.
**Solutions**:
```bash
# Option 1: Disable API validation (faster deployment)
kubectl apply -f nginx-nodeport-configmaps.yaml --validate=false
kubectl apply -f nginx-nodeport-deployment.yaml --validate=false
kubectl apply -f nginx-nodeport-service.yaml --validate=false
# Option 2: Use client-side dry-run to verify YAML syntax
kubectl apply -f nginx-nodeport-configmaps.yaml --dry-run=client
kubectl apply -f nginx-nodeport-deployment.yaml --dry-run=client
kubectl apply -f nginx-nodeport-service.yaml --dry-run=client
# Option 3: Increase timeout
kubectl apply -f nginx-nodeport-configmaps.yaml --timeout=5m
```
### Issue 2: Network Connectivity Problems
**Error**: `Unable to connect to the server`
**Diagnose and Fix**:
```bash
# Check cluster connectivity
kubectl cluster-info
# Check node status
kubectl get nodes
# Verify kubeconfig
kubectl config view
# Test basic connectivity
kubectl get pods --all-namespaces
```
### Issue 3: Slow API Responses
**Error**: Operations take very long or timeout
**Performance Optimizations**:
```bash
# Use smaller output formats
kubectl get pods -l app=nginx-nodeport -o wide
# Disable unnecessary features
kubectl apply -f nginx-nodeport-configmaps.yaml --v=1
# Use specific resource targeting
kubectl apply -f nginx-nodeport-configmaps.yaml --namespace=default
```
## 🔍 Pre-Deployment Validation
### YAML Syntax Validation (No Cluster Required)
```bash
# Validate all files without cluster connection
for file in *.yaml; do
echo "Validating $file..."
kubectl create -f "$file" --dry-run=client --validate=false
done
```
### File-by-File Validation
```bash
# Test each component individually
echo "=== Testing ConfigMaps ==="
kubectl apply -f nginx-nodeport-configmaps.yaml --dry-run=client
echo "=== Testing Deployment ==="
kubectl apply -f nginx-nodeport-deployment.yaml --dry-run=client
echo "=== Testing Service ==="
kubectl apply -f nginx-nodeport-service.yaml --dry-run=client
```
## 🚀 Alternative Deployment Methods
### Method 1: Manual Resource Creation
If `kubectl apply` fails, create resources manually:
```bash
# 1. Create ConfigMaps
kubectl create configmap nginx-nodeport-content --from-literal=index.html="<h1>Hello NodePort</h1>"
kubectl create configmap nginx-nodeport-nginx-config --from-literal=default.conf="server { listen 8080; }"
# 2. Create Deployment
kubectl create deployment nginx-nodeport --image=nginx:alpine --replicas=1
# 3. Expose Service
kubectl expose deployment nginx-nodeport --type=NodePort --port=8080 --target-port=8080
```
### Method 2: Using Helm (Advanced)
Create a Helm chart for easier deployment:
```bash
# Install Helm if not present
# Then package and deploy
helm package .
helm install nginx-nodeport ./nginx-nodeport-0.1.0.tgz
```
### Method 3: Scripted Deployment
Use the provided deployment script:
```bash
# Make script executable
chmod +x deploy-nginx-nodeport.sh
# Run automated deployment
./deploy-nginx-nodeport.sh
```
## 🛠️ Creating deployment-nginx-nodeport.sh
Let me create a robust deployment script that handles connectivity issues:
```bash
#!/bin/bash
# nginx-nodeport Deployment Script
# Handles connectivity issues and provides fallback options
set -e
echo "🚀 Deploying nginx-nodeport to Mycelium Cloud"
echo "=============================================="
# Colors
GREEN='\033[0;32m'
YELLOW='\033[1;33m'
RED='\033[0;31m'
NC='\033[0m'
print_success() { echo -e "${GREEN}$1${NC}"; }
print_warning() { echo -e "${YELLOW}⚠️ $1${NC}"; }
print_error() { echo -e "${RED}$1${NC}"; }
# Function to deploy with fallback options
deploy_with_fallback() {
local file=$1
local name=$2
echo "📋 Deploying $name..."
# Try normal deployment
if kubectl apply -f "$file" --validate=false 2>/dev/null; then
print_success "$name deployed successfully"
return 0
fi
# Try with client validation only
print_warning "Normal deployment failed, trying client-side validation..."
if kubectl apply -f "$file" --dry-run=client --validate=false 2>/dev/null; then
print_warning "$name validated (client-side only) - cluster may be unavailable"
return 1
fi
# Try manual creation for specific resource types
print_error "$name deployment failed completely"
return 2
}
# Check cluster connectivity
echo "🔍 Checking cluster connectivity..."
if kubectl cluster-info &>/dev/null; then
print_success "Cluster is accessible"
CLUSTER_AVAILABLE=true
else
print_warning "Cluster is not accessible - will use client-side validation only"
CLUSTER_AVAILABLE=false
fi
# Deploy ConfigMaps
if [ "$CLUSTER_AVAILABLE" = true ]; then
deploy_with_fallback "nginx-nodeport-configmaps.yaml" "ConfigMaps"
else
echo "🔍 Validating ConfigMaps (client-side)..."
if kubectl create -f nginx-nodeport-configmaps.yaml --dry-run=client --validate=false &>/dev/null; then
print_success "ConfigMaps validated (client-side)"
else
print_error "ConfigMaps validation failed"
exit 1
fi
fi
# Deploy Deployment
if [ "$CLUSTER_AVAILABLE" = true ]; then
deploy_with_fallback "nginx-nodeport-deployment.yaml" "Deployment"
else
echo "🔍 Validating Deployment (client-side)..."
if kubectl create -f nginx-nodeport-deployment.yaml --dry-run=client --validate=false &>/dev/null; then
print_success "Deployment validated (client-side)"
else
print_error "Deployment validation failed"
exit 1
fi
fi
# Deploy Service
if [ "$CLUSTER_AVAILABLE" = true ]; then
deploy_with_fallback "nginx-nodeport-service.yaml" "Service"
else
echo "🔍 Validating Service (client-side)..."
if kubectl create -f nginx-nodeport-service.yaml --dry-run=client --validate=false &>/dev/null; then
print_success "Service validated (client-side)"
else
print_error "Service validation failed"
exit 1
fi
fi
# Final status
if [ "$CLUSTER_AVAILABLE" = true ]; then
echo ""
echo "🎉 Deployment Complete!"
echo "======================"
echo "📊 Checking deployment status..."
kubectl get all -l app=nginx-nodeport
# Get access information
NODE_IPV6=$(kubectl get nodes -o jsonpath='{.items[0].status.addresses[?(@.type=="InternalIP")].address}' 2>/dev/null || echo "YOUR-NODE-IPV6")
echo ""
echo "🌐 Access your website at: http://[$NODE_IPV6]:30090"
echo "🧪 Run health check: curl -6 http://[$NODE_IPV6]:30090/health"
echo "📋 Test script: ./test-nodeport-ipv6.sh"
else
echo ""
print_warning "Deployment files are valid but cluster is not accessible"
echo "📝 When cluster becomes available, run:"
echo " kubectl apply -f nginx-nodeport-configmaps.yaml --validate=false"
echo " kubectl apply -f nginx-nodeport-deployment.yaml --validate=false"
echo " kubectl apply -f nginx-nodeport-service.yaml --validate=false"
fi
echo ""
echo "📚 Next Steps:"
echo " • Check documentation: nginx-nodeport.md"
echo " • Compare approaches: compare-approaches.md"
echo " • Run validation: deployment-validation.md"
```
## 🧪 Testing Without Cluster
### Syntax Validation
```bash
# Test all YAML files for syntax errors
for file in *.yaml; do
echo "Testing $file syntax..."
python3 -c "
import yaml
import sys
try:
with open('$file') as f:
yaml.safe_load_all(f)
print('✅ $file: Valid YAML')
except Exception as e:
print('❌ $file: Invalid -', str(e))
sys.exit(1)
"
done
```
### Lint Kubernetes YAML
```bash
# Install kube-linter if not present
# wget https://github.com/stackrox/kube-linter/releases/download/v0.6.0/kube-linter_Linux_x86_64.tar.gz
# tar xzf kube-linter_Linux_x86_64.tar.gz
# sudo mv kube-linter /usr/local/bin/
# Lint all files
kube-linter lint *.yaml
```
## 🔧 Cluster-Specific Issues
### Mycelium Cloud Specific
```bash
# Check Mycelium-specific connectivity
kubectl get nodes -o wide
kubectl describe node | grep -A 5 -B 5 "Addresses"
# Test IPv6 connectivity
ping6 -c 3 $(kubectl get nodes -o jsonpath='{.items[0].status.addresses[?(@.type=="InternalIP")].address}')
# Check Mycelium network plugin
kubectl get pods -n kube-system | grep mycelium
```
### Local Development Clusters
```bash
# For minikube
minikube status
minikube start --driver=docker
# For kind
kind get clusters
kind create cluster --name mycelium-test
# For k3s
sudo systemctl status k3s
```
## 📊 Performance Optimization
### Faster Operations
```bash
# Use shorter timeouts for faster failure
kubectl get pods --timeout=30s
# Reduce verbosity
kubectl get pods -l app=nginx-nodeport -o json
# Use specific namespaces
kubectl config set-context --current --namespace=default
```
### Resource-Afficient Queries
```bash
# Lightweight status checks
kubectl get pods -l app=nginx-nodeport --no-headers
# Minimal output
kubectl get svc nginx-nodeport-service -o name
# Batch operations
kubectl get deployment,svc,configmap -l app=nginx-nodeport
```
## 🆘 Emergency Deployment
If all else fails, here's a minimal emergency deployment:
```bash
# Emergency nginx deployment
kubectl create deployment emergency-nginx --image=nginx:alpine
kubectl expose deployment emergency-nginx --type=NodePort --port=80
kubectl patch service emergency-nginx -p '{"spec":{"ports":[{"port":80,"targetPort":80,"nodePort":30090}]}}'
# Get access
NODE_IP=$(kubectl get nodes -o jsonpath='{.items[0].status.addresses[?(@.type=="InternalIP")].address}')
echo "Emergency site: http://$NODE_IP:30090"
```
This comprehensive troubleshooting guide should help resolve most deployment issues and provide multiple fallback options for different scenarios.