Fix multi-server TLS setup to support password authentication and improve error handling

This commit is contained in:
Kelin
2026-01-26 18:16:35 -05:00
parent e2d28b5208
commit 62454130db
2 changed files with 116 additions and 174 deletions

View File

@@ -190,7 +190,7 @@ This section provides a complete deployment plan for scenarios where the core in
## Prerequisites
- Both servers must be on the same network and able to communicate
- SSH access configured between servers (passwordless recommended for automation)
- SSH access configured between servers (key-based or password authentication supported)
- Domain configured with DuckDNS or similar
- The EZ-Homelab script handles most Docker TLS and certificate setup automatically
- Basic understanding of Docker concepts (optional - script guides you through setup)

View File

@@ -71,42 +71,131 @@ generate_shared_ca() {
# Function to setup multi-server TLS for remote servers
setup_multi_server_tls() {
local ca_dir="/opt/stacks/core/shared-ca"
sudo mkdir -p "$ca_dir"
sudo chown "$ACTUAL_USER:$ACTUAL_USER" "$ca_dir"
if [ -n "$CORE_SERVER_IP" ]; then
log_info "Setting up multi-server TLS using shared CA from core server $CORE_SERVER_IP..."
else
# Prompt for core server IP if not set
read -p "Enter the IP address of your core server: " CORE_SERVER_IP
while [ -z "$CORE_SERVER_IP" ]; do
log_warning "Core server IP is required for shared TLS"
read -p "Enter the IP address of your core server: " CORE_SERVER_IP
done
log_info "Setting up multi-server TLS using shared CA from core server $CORE_SERVER_IP..."
fi
# Create shared-ca directory
local ca_dir="/opt/stacks/core/shared-ca"
sudo mkdir -p "$ca_dir"
sudo chown "$ACTUAL_USER:$ACTUAL_USER" "$ca_dir"
# Prompt for SSH username if not set
if [ -z "$SSH_USER" ]; then
DEFAULT_SSH_USER="${DEFAULT_USER:-$USER}"
read -p "SSH username for core server [$DEFAULT_SSH_USER]: " SSH_USER
SSH_USER="${SSH_USER:-$DEFAULT_SSH_USER}"
fi
# Copy shared CA from core server
if scp -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null "$ACTUAL_USER@$CORE_SERVER_IP:/opt/stacks/core/shared-ca/ca.pem" "$ca_dir/" 2>/dev/null; then
log_success "Shared CA copied from core server"
# Test SSH connection - try key authentication first
log_info "Testing SSH connection to core server ($SSH_USER@$CORE_SERVER_IP)..."
if ssh -o ConnectTimeout=10 -o StrictHostKeyChecking=no -o BatchMode=yes "$SSH_USER@$CORE_SERVER_IP" "echo 'SSH connection successful'" 2>/dev/null; then
log_success "SSH connection established using key authentication"
USE_SSHPASS=false
else
# Key authentication failed, try password authentication
log_info "Key authentication failed, trying password authentication..."
read -s -p "Enter SSH password for $SSH_USER@$CORE_SERVER_IP: " SSH_PASSWORD
echo ""
if sshpass -p "$SSH_PASSWORD" ssh -o ConnectTimeout=10 -o StrictHostKeyChecking=no "$SSH_USER@$CORE_SERVER_IP" "echo 'SSH connection successful'" 2>/dev/null; then
log_success "SSH connection established using password authentication"
USE_SSHPASS=true
else
log_warning "Failed to copy shared CA from core server $CORE_SERVER_IP"
log_warning "This will create TLS certificate mismatches between core and remote servers"
log_warning "Sablier will not be able to connect to this remote Docker daemon"
TLS_ISSUES_SUMMARY="⚠️ TLS Configuration Issue: Could not copy shared CA from core server $CORE_SERVER_IP
log_error "Cannot connect to core server via SSH. Please check:"
echo " 1. SSH is running on the core server"
echo " 2. SSH keys are properly configured, or username/password are correct"
echo " 3. The core server IP is correct"
echo ""
TLS_ISSUES_SUMMARY="⚠️ TLS Configuration Issue: Cannot connect to core server $CORE_SERVER_IP via SSH
This will prevent Sablier from connecting to remote Docker daemons.
To fix this:
1. Ensure SSH access works: ssh $ACTUAL_USER@$CORE_SERVER_IP
2. Verify core server has: /opt/stacks/core/shared-ca/ca.pem
3. Manually copy CA: scp $ACTUAL_USER@$CORE_SERVER_IP:/opt/stacks/core/shared-ca/ca.pem $ca_dir/
4. Regenerate server certificates on this server with shared CA
5. Restart Docker: sudo systemctl restart docker
6. Test connection: docker --tlsverify --tlscacert=$ca_dir/ca.pem --tlscert=$ca_dir/cert.pem --tlskey=$ca_dir/key.pem --host=tcp://localhost:2376 ps
1. Ensure SSH is running on the core server
2. Configure SSH keys or provide correct password
3. Verify the core server IP is correct
4. Test SSH connection: ssh $SSH_USER@$CORE_SERVER_IP
Without SSH access, shared CA cannot be fetched for secure multi-server TLS."
return
fi
fi
# Fetch shared CA certificates from core server
log_info "Fetching shared CA certificates from core server..."
SHARED_CA_EXISTS=false
# Check if shared CA exists on core server (check both old and new locations)
if [ "$USE_SSHPASS" = true ] && [ -n "$SSH_PASSWORD" ]; then
if sshpass -p "$SSH_PASSWORD" ssh -o StrictHostKeyChecking=no "$SSH_USER@$CORE_SERVER_IP" "[ -f /opt/stacks/core/shared-ca/ca.pem ] && [ -f /opt/stacks/core/shared-ca/ca-key.pem ] && [ -r /opt/stacks/core/shared-ca/ca.pem ] && [ -r /opt/stacks/core/shared-ca/ca-key.pem ]" 2>/dev/null; then
SHARED_CA_EXISTS=true
SHARED_CA_PATH="/opt/stacks/core/shared-ca"
log_info "Detected CA certificate and key in shared-ca location"
elif sshpass -p "$SSH_PASSWORD" ssh -o StrictHostKeyChecking=no "$SSH_USER@$CORE_SERVER_IP" "[ -f /opt/stacks/core/docker-tls/ca.pem ] && [ -f /opt/stacks/core/docker-tls/ca-key.pem ] && [ -r /opt/stacks/core/docker-tls/ca.pem ] && [ -r /opt/stacks/core/docker-tls/ca-key.pem ]" 2>/dev/null; then
SHARED_CA_EXISTS=true
SHARED_CA_PATH="/opt/stacks/core/docker-tls"
log_info "Detected CA certificate and key in docker-tls location"
fi
else
if ssh -o StrictHostKeyChecking=no "$SSH_USER@$CORE_SERVER_IP" "[ -f /opt/stacks/core/shared-ca/ca.pem ] && [ -f /opt/stacks/core/shared-ca/ca-key.pem ] && [ -r /opt/stacks/core/shared-ca/ca.pem ] && [ -r /opt/stacks/core/shared-ca/ca-key.pem ]" 2>/dev/null; then
SHARED_CA_EXISTS=true
SHARED_CA_PATH="/opt/stacks/core/shared-ca"
log_info "Detected CA certificate and key in shared-ca location"
elif ssh -o StrictHostKeyChecking=no "$SSH_USER@$CORE_SERVER_IP" "[ -f /opt/stacks/core/docker-tls/ca.pem ] && [ -f /opt/stacks/core/docker-tls/ca-key.pem ] && [ -r /opt/stacks/core/docker-tls/ca.pem ] && [ -r /opt/stacks/core/docker-tls/ca-key.pem ]" 2>/dev/null; then
SHARED_CA_EXISTS=true
SHARED_CA_PATH="/opt/stacks/core/docker-tls"
log_info "Detected CA certificate and key in docker-tls location"
fi
fi
if [ "$SHARED_CA_EXISTS" = true ]; then
# Copy existing shared CA from core server
set +e
scp_output=$(scp -o StrictHostKeyChecking=no "$SSH_USER@$CORE_SERVER_IP:$SHARED_CA_PATH/ca.pem" "$SSH_USER@$CORE_SERVER_IP:$SHARED_CA_PATH/ca-key.pem" "$ca_dir/" 2>&1)
scp_exit_code=$?
set -e
if [ $scp_exit_code -eq 0 ]; then
log_success "Shared CA certificate and key fetched from core server"
setup_docker_tls
else
log_error "Failed to fetch shared CA certificate and key from core server"
TLS_ISSUES_SUMMARY="⚠️ TLS Configuration Issue: Could not copy shared CA from core server $CORE_SERVER_IP
SCP Error: $scp_output
To fix this:
1. Ensure SSH key authentication works: ssh $ACTUAL_USER@$CORE_SERVER_IP
2. Verify core server has: $SHARED_CA_PATH/ca.pem and ca-key.pem
3. Check file permissions on core server: ls -la $SHARED_CA_PATH/
4. Manually copy CA: scp $ACTUAL_USER@$CORE_SERVER_IP:$SHARED_CA_PATH/ca.pem $ca_dir/
scp $ACTUAL_USER@$CORE_SERVER_IP:$SHARED_CA_PATH/ca-key.pem $ca_dir/
5. Regenerate server certificates: run setup_docker_tls after copying
6. Restart Docker: sudo systemctl restart docker
Then restart Sablier on the core server to reconnect."
generate_shared_ca
return
fi
# Setup Docker TLS with shared CA
setup_docker_tls
else
log_info "No core server specified, setting up local TLS..."
generate_shared_ca
setup_docker_tls
log_warning "Shared CA certificates not found on core server."
log_info "Please ensure the core server has been set up first and has generated the shared CA certificates."
TLS_ISSUES_SUMMARY="⚠️ TLS Configuration Issue: Shared CA certificates not found on core server $CORE_SERVER_IP
This will prevent Sablier from connecting to remote Docker daemons.
To fix this:
1. Ensure the core server is set up and has generated shared CA certificates
2. Verify SSH access: ssh $ACTUAL_USER@$CORE_SERVER_IP
3. Check core server locations: /opt/stacks/core/shared-ca/ or /opt/stacks/core/docker-tls/
4. Manually copy CA certificates if needed
5. Re-run the infrastructure deployment
Without shared CA, remote Docker access will not work securely."
return
fi
}
@@ -336,148 +425,6 @@ prompt_for_values() {
echo ""
}
# Certificate fetching function for infrastructure-only deployments
get_certs_from_core_server() {
log_info "Infrastructure-only deployment detected. Fetching certificate from core server for remote Docker control..."
# Prompt for core server IP
read -p "Enter the IP address of your core server: " CORE_SERVER_IP
while [ -z "$CORE_SERVER_IP" ]; do
log_warning "Core server IP is required for certificate fetching"
read -p "Enter the IP address of your core server: " CORE_SERVER_IP
done
# Prompt for SSH username
DEFAULT_SSH_USER="${DEFAULT_USER:-$USER}"
read -p "SSH username for core server [$DEFAULT_SSH_USER]: " SSH_USER
SSH_USER="${SSH_USER:-$DEFAULT_SSH_USER}"
# Test SSH connection - try key authentication first
log_info "Testing SSH connection to core server ($SSH_USER@$CORE_SERVER_IP)..."
if ssh -o ConnectTimeout=10 -o StrictHostKeyChecking=no -o BatchMode=yes "$SSH_USER@$CORE_SERVER_IP" "echo 'SSH connection successful'" 2>/dev/null; then
log_success "SSH connection established using key authentication"
USE_SSHPASS=false
else
# Key authentication failed, try password authentication
log_info "Key authentication failed, trying password authentication..."
read -s -p "Enter SSH password for $SSH_USER@$CORE_SERVER_IP: " SSH_PASSWORD
echo ""
if sshpass -p "$SSH_PASSWORD" ssh -o ConnectTimeout=10 -o StrictHostKeyChecking=no "$SSH_USER@$CORE_SERVER_IP" "echo 'SSH connection successful'" 2>/dev/null; then
log_success "SSH connection established using password authentication"
USE_SSHPASS=true
else
log_error "Cannot connect to core server via SSH. Please check:"
echo " 1. SSH is running on the core server"
echo " 2. SSH keys are properly configured, or username/password are correct"
echo " 3. The core server IP is correct"
echo ""
read -p "Do you want to continue anyway? (y/N): " -n 1 -r
echo ""
if [[ ! $REPLY =~ ^[Yy]$ ]]; then
log_error "Certificate fetching cancelled. Please verify SSH access and try again."
exit 1
fi
USE_SSHPASS=true # Assume password auth for copying
fi
fi
# Fetch shared CA certificates from core server
log_info "Fetching shared CA certificates from core server..."
sudo mkdir -p "/opt/stacks/core/shared-ca"
# First check if shared CA exists on core server (check both old and new locations)
SHARED_CA_EXISTS=false
if [ "$USE_SSHPASS" = true ] && [ -n "$SSH_PASSWORD" ]; then
if sshpass -p "$SSH_PASSWORD" ssh -o StrictHostKeyChecking=no "$SSH_USER@$CORE_SERVER_IP" "[ -f /opt/stacks/core/shared-ca/ca.pem ] && [ -f /opt/stacks/core/shared-ca/ca-key.pem ] && [ -r /opt/stacks/core/shared-ca/ca.pem ] && [ -r /opt/stacks/core/shared-ca/ca-key.pem ]" 2>/dev/null; then
SHARED_CA_EXISTS=true
SHARED_CA_PATH="/opt/stacks/core/shared-ca"
log_info "Detected CA certificate and key in shared-ca location"
elif sshpass -p "$SSH_PASSWORD" ssh -o StrictHostKeyChecking=no "$SSH_USER@$CORE_SERVER_IP" "[ -f /opt/stacks/core/docker-tls/ca.pem ] && [ -f /opt/stacks/core/docker-tls/ca-key.pem ] && [ -r /opt/stacks/core/docker-tls/ca.pem ] && [ -r /opt/stacks/core/docker-tls/ca-key.pem ]" 2>/dev/null; then
SHARED_CA_EXISTS=true
SHARED_CA_PATH="/opt/stacks/core/docker-tls"
log_info "Detected CA certificate and key in docker-tls location"
fi
else
if ssh -o StrictHostKeyChecking=no "$SSH_USER@$CORE_SERVER_IP" "[ -f /opt/stacks/core/shared-ca/ca.pem ] && [ -f /opt/stacks/core/shared-ca/ca-key.pem ] && [ -r /opt/stacks/core/shared-ca/ca.pem ] && [ -r /opt/stacks/core/shared-ca/ca-key.pem ]" 2>/dev/null; then
SHARED_CA_EXISTS=true
SHARED_CA_PATH="/opt/stacks/core/shared-ca"
log_info "Detected CA certificate and key in shared-ca location"
elif ssh -o StrictHostKeyChecking=no "$SSH_USER@$CORE_SERVER_IP" "[ -f /opt/stacks/core/docker-tls/ca.pem ] && [ -f /opt/stacks/core/docker-tls/ca-key.pem ] && [ -r /opt/stacks/core/docker-tls/ca.pem ] && [ -r /opt/stacks/core/docker-tls/ca-key.pem ]" 2>/dev/null; then
SHARED_CA_EXISTS=true
SHARED_CA_PATH="/opt/stacks/core/docker-tls"
log_info "Detected CA certificate and key in docker-tls location"
fi
fi
if [ "$SHARED_CA_EXISTS" = true ]; then
# Copy existing shared CA from core server
if [ "$USE_SSHPASS" = true ] && [ -n "$SSH_PASSWORD" ]; then
log_info "Running: sshpass -p [PASSWORD] scp -o StrictHostKeyChecking=no $SSH_USER@$CORE_SERVER_IP:$SHARED_CA_PATH/ca.pem $SSH_USER@$CORE_SERVER_IP:$SHARED_CA_PATH/ca-key.pem /opt/stacks/core/shared-ca/"
if sshpass -p "$SSH_PASSWORD" scp -o StrictHostKeyChecking=no "$SSH_USER@$CORE_SERVER_IP:$SHARED_CA_PATH/ca.pem" "$SSH_USER@$CORE_SERVER_IP:$SHARED_CA_PATH/ca-key.pem" "/opt/stacks/core/shared-ca/" 2>&1; then
log_success "Shared CA certificate and key fetched from core server"
# Generate server certificates signed by the shared CA
setup_docker_tls
else
log_warning "Failed to fetch shared CA certificate and key from core server"
SHARED_CA_EXISTS=false
fi
else
log_info "Running: scp -o StrictHostKeyChecking=no $SSH_USER@$CORE_SERVER_IP:$SHARED_CA_PATH/ca.pem $SSH_USER@$CORE_SERVER_IP:$SHARED_CA_PATH/ca-key.pem /opt/stacks/core/shared-ca/"
if scp -o StrictHostKeyChecking=no "$SSH_USER@$CORE_SERVER_IP:$SHARED_CA_PATH/ca.pem" "$SSH_USER@$CORE_SERVER_IP:$SHARED_CA_PATH/ca-key.pem" "/opt/stacks/core/shared-ca/" 2>&1; then
log_success "Shared CA certificate and key fetched from core server"
# Generate server certificates signed by the shared CA
setup_docker_tls
else
log_warning "Failed to fetch shared CA certificate and key from core server"
SHARED_CA_EXISTS=false
fi
fi
fi
if [ "$SHARED_CA_EXISTS" = false ]; then
log_warning "Shared CA certificates not found on core server."
log_info "Please ensure the core server has been set up first and has generated the shared CA certificates."
log_info "You can manually copy the certificates later (check both locations on core server):"
echo " sudo mkdir -p /opt/stacks/core/shared-ca"
echo " # Try shared-ca location:"
echo " sudo scp $SSH_USER@$CORE_SERVER_IP:/opt/stacks/core/shared-ca/ca.pem /opt/stacks/core/shared-ca/"
echo " sudo scp $SSH_USER@$CORE_SERVER_IP:/opt/stacks/core/shared-ca/ca-key.pem /opt/stacks/core/shared-ca/"
echo " # Or docker-tls location:"
echo " sudo scp $SSH_USER@$CORE_SERVER_IP:/opt/stacks/core/docker-tls/ca.pem /opt/stacks/core/shared-ca/"
echo " sudo scp $SSH_USER@$CORE_SERVER_IP:/opt/stacks/core/docker-tls/ca-key.pem /opt/stacks/core/shared-ca/"
echo " sudo chown -R $SSH_USER:$SSH_USER /opt/stacks/core/shared-ca"
echo ""
log_info "Then restart the Docker daemon: sudo systemctl reload docker"
echo ""
fi
# Update Docker daemon configuration to use shared CA
if [ "$SHARED_CA_EXISTS" = true ]; then
log_info "Updating Docker daemon to use shared CA for TLS..."
if [ -f "/opt/stacks/core/shared-ca/ca.pem" ]; then
# Update daemon.json to use the shared CA for both server and client verification
cat > /tmp/daemon.json <<EOF
{
"tls": true,
"tlsverify": true,
"tlscacert": "/opt/stacks/core/shared-ca/ca.pem",
"tlscert": "/home/$USER/EZ-Homelab/docker-tls/server-cert.pem",
"tlskey": "/home/$USER/EZ-Homelab/docker-tls/server-key.pem"
}
EOF
sudo cp /tmp/daemon.json /etc/docker/daemon.json
sudo systemctl restart docker
log_success "Docker daemon updated to use shared CA"
log_info "Core server can now securely connect to this Docker daemon using shared CA"
else
log_warning "Shared CA certificate not found, daemon configuration not updated"
fi
fi
echo ""
}
# System setup function (Docker, directories, etc.)
system_setup() {
log_info "Performing system setup..."
@@ -1037,11 +984,6 @@ main() {
# Save configuration
save_env_file
# Handle certificate sharing for infrastructure-only deployments
if [ "$MAIN_CHOICE" = "3" ]; then
get_certs_from_core_server
fi
# Perform deployment
perform_deployment