Files
fleet-demo/DNS-SERVER-GUIDE.md

221 lines
5.0 KiB
Markdown

# Multi-Tenant DNS Server Setup
## 🎯 Overview
This DNS server provides:
- **External DNS** for connectvm.cloud (public access)
- **Multi-tenant DNS** for dev, prod, and QA teams
- **High Availability** with 3 replicas
- **Metrics** for monitoring
## 🌐 DNS Zones Configured
### 1. Main Domain: `connectvm.cloud`
**Public DNS for main services**
Available records:
- `rancher.connectvm.cloud` → 160.30.114.10
- `paste.connectvm.cloud` → 160.30.114.10
- `fleet.connectvm.cloud` → 160.30.114.10
- `hello.connectvm.cloud` → 160.30.114.10
- `dns.connectvm.cloud` → 160.30.114.10
- `*.connectvm.cloud` → 160.30.114.10 (wildcard)
### 2. Dev Team: `dev.connectvm.cloud`
**Development team's DNS zone**
Available records:
- `app1.dev.connectvm.cloud`
- `app2.dev.connectvm.cloud`
- `api.dev.connectvm.cloud`
- `web.dev.connectvm.cloud`
- `dashboard.dev.connectvm.cloud`
- `jenkins.dev.connectvm.cloud`
- `gitlab.dev.connectvm.cloud`
- `db.dev.connectvm.cloud`
- `redis.dev.connectvm.cloud`
- `*.dev.connectvm.cloud` (wildcard)
### 3. Prod Team: `prod.connectvm.cloud`
**Production team's DNS zone**
Available records:
- `api.prod.connectvm.cloud`
- `web.prod.connectvm.cloud`
- `app.prod.connectvm.cloud`
- `admin.prod.connectvm.cloud`
- `portal.prod.connectvm.cloud`
- `db.prod.connectvm.cloud`
- `monitoring.prod.connectvm.cloud`
- `*.prod.connectvm.cloud` (wildcard)
### 4. QA Team: `qa.connectvm.cloud`
**QA/Testing team's DNS zone**
Available records:
- `test.qa.connectvm.cloud`
- `staging.qa.connectvm.cloud`
- `selenium.qa.connectvm.cloud`
- `automation.qa.connectvm.cloud`
- `reports.qa.connectvm.cloud`
- `*.qa.connectvm.cloud` (wildcard)
## 🔧 External Access
### DNS Server IPs:
- **Primary**: `160.30.114.10:30053` (UDP/TCP)
- **Internal**: `10.96.100.100:53` (ClusterIP)
### Configure Clients to Use DNS:
**Linux/Mac:**
```bash
# Add to /etc/resolv.conf
nameserver 160.30.114.10
# Or use specific port with dig:
dig @160.30.114.10 -p 30053 rancher.connectvm.cloud
```
**Windows:**
```powershell
# Set DNS server
netsh interface ip set dns "Ethernet" static 160.30.114.10
```
**Kubernetes Pods:**
```yaml
spec:
dnsPolicy: None
dnsConfig:
nameservers:
- 10.96.100.100
searches:
- connectvm.cloud
- dev.connectvm.cloud
```
## 📊 Monitoring
Access DNS metrics:
- **URL**: https://dns.connectvm.cloud/metrics
- **Prometheus endpoint**: Port 9153
Metrics include:
- Query count
- Query types
- Response codes
- Cache hit/miss rates
- Zone transfer stats
## 🔒 Security Features
- **Zone isolation**: Each tenant has separate DNS zone
- **No zone transfers**: Zones are read-only
- **Query logging**: All queries are logged
- **Rate limiting**: Built-in caching (300s TTL)
## 🛠️ Management
### Add New Record:
1. Edit `dns-server.yaml` ConfigMap
2. Add record to appropriate zone file
3. Increment Serial number
4. Git push → Fleet auto-deploys
**Example** - Add new dev app:
```
newapp IN A 160.30.114.10
```
### Add New Tenant Zone:
1. Create new zone file in ConfigMap
2. Add zone to Corefile
3. Git push
## 🧪 Testing
Test DNS resolution:
```bash
# Test main domain
dig @160.30.114.10 -p 30053 rancher.connectvm.cloud
# Test dev tenant
dig @160.30.114.10 -p 30053 app1.dev.connectvm.cloud
# Test prod tenant
dig @160.30.114.10 -p 30053 api.prod.connectvm.cloud
# Test QA tenant
dig @160.30.114.10 -p 30053 test.qa.connectvm.cloud
# Test wildcard
dig @160.30.114.10 -p 30053 anything.dev.connectvm.cloud
```
## 📱 Use Cases
### For Dev Team:
```bash
# Deploy app with custom DNS
kubectl create deployment myapp --image=nginx
kubectl expose deployment myapp --port=80
# Access via: myapp.dev.connectvm.cloud
```
### For Prod Team:
```bash
# Production API endpoint
api.prod.connectvm.cloud → Production API server
```
### For QA Team:
```bash
# Automated testing
selenium.qa.connectvm.cloud → Selenium Grid
automation.qa.connectvm.cloud → Test Runner
```
## 🚀 High Availability
- **3 replicas** across different nodes
- **Anti-affinity** rules for pod distribution
- **Auto-restart** on failure
- **Health checks** every 10 seconds
## 📝 DNS Server Info
- **Software**: CoreDNS 1.11.1
- **Protocol**: DNS (UDP/TCP port 53)
- **Upstream**: Google DNS (8.8.8.8, 8.8.4.4)
- **Cache TTL**: 300 seconds
- **Zone TTL**: 3600 seconds (1 hour)
## 🎯 Architecture
```
External Queries (port 30053)
NodePort Service
DNS Pods (3 replicas)
┌─────────────┬──────────────┬─────────────┐
│ Main Zone │ Dev Zone │ Prod Zone │
│ qa Zone │ K8s DNS │ Upstream │
└─────────────┴──────────────┴─────────────┘
```
## 🔄 Updates
All DNS updates happen via GitOps:
1. Edit zone files in Git
2. Push to Gitea
3. Fleet auto-deploys in ~15 seconds
4. DNS records updated automatically
No manual DNS server management needed!