CSTF/wild-cloud
1
1
Fork 0
Code Issues 12 Pull Requests Actions Packages Projects 1 Releases 1 Wiki Activity

Refactor wild-cluster-config-generate script; remove --force option and improve cluster secrets generation. Add wild-cluster-node-boot-assets-download script for downloading Talos installer images and assets. Update wild-setup-cluster to integrate new boot asset download process and adjust configuration steps. Clean up wild-setup-services and update documentation for clarity on new processes.

Browse Source
This commit is contained in:
Paul Payne
2025-07-06 14:34:58 -07:00
parent d1912af913
commit 93240f6300
8 changed files with 190 additions and 251 deletions
Download Patch File Download Diff File Expand all files Collapse all files

227
bin/wild-setup-cluster
View File

@@ -66,22 +66,31 @@ else
init_wild_env
fi
# Check basic configuration
check_basic_config
print_header "Wild Cloud Cluster Setup"
print_info "Setting up cluster infrastructure"
echo ""
# =============================================================================
# Generate initial cluster configuration
wild-cluster-config-generate
# Configure Talos cli with our new cluster context
HAS_CONTEXT=$(talos config get | grep -c "<cluster.name>")
if [ "$HAS_CONTEXT" -eq 0 ]; then
print_info "No Talos context found for cluster <cluster.name>, creating..."
talos config merge ${WC_HOME}/setup/cluster-nodes/generated/talosconfig
talos config use <cluster.name>
print_success "Talos context for <cluster.name> created and set as current"
fi
# Talos asset download
# =============================================================================
if [ "${SKIP_INSTALLER}" = false ]; then
print_header "Installer Image Generation"
print_info "Running wild-cluster-node-image-create..."
wild-cluster-node-image-create
print_info "Running wild-cluster-node-boot-assets-download..."
wild-cluster-node-boot-assets-download
print_success "Installer image generated"
echo ""
@@ -90,36 +99,21 @@ else
fi
# =============================================================================
# Network and Cluster Configuration
# Configuration
# =============================================================================
prompt_if_unset_config "owner.email" "Owner email address"
# Configure network settings
if [ -z "$(get_current_config "cloud.router.ip")" ]; then
print_header "Network Configuration"
CURRENT_IP=$(ip route get 8.8.8.8 | awk '{print $7; exit}' 2>/dev/null || echo "192.168.1.100")
GATEWAY_IP=$(ip route | grep default | awk '{print $3; exit}' 2>/dev/null || echo "192.168.1.1")
SUBNET_PREFIX=$(echo "${CURRENT_IP}" | cut -d. -f1-3)
prompt_if_unset_config "cloud.router.ip" "Router/Gateway IP" "${GATEWAY_IP}"
prompt_if_unset_config "cloud.dns.ip" "DNS server IP (dnsmasq machine)" "${SUBNET_PREFIX}.50"
prompt_if_unset_config "cloud.dhcpRange" "DHCP range for dnsmasq" "${SUBNET_PREFIX}.100,${SUBNET_PREFIX}.200"
prompt_if_unset_config "cloud.dnsmasq.interface" "Network interface for dnsmasq" "eth0"
prompt_if_unset_config "cloud.dns.externalResolver" "External DNS resolver" "1.1.1.1"
print_success "Network configuration completed"
echo ""
fi
# Configure cluster settings
print_header "Kubernetes Cluster Configuration"
CURRENT_IP=$(ip route get 8.8.8.8 | awk '{print $7; exit}' 2>/dev/null || echo "192.168.1.100")
GATEWAY_IP=$(ip route | grep default | awk '{print $3; exit}' 2>/dev/null || echo "192.168.1.1")
SUBNET_PREFIX=$(echo "${CURRENT_IP}" | cut -d. -f1-3)
# Talos version
prompt_if_unset_config "cluster.nodes.talos.version" "Talos version" "v1.10.4"
talos_version=$(wild-config "cluster.nodes.talos.version")
prompt_if_unset_config "cloud.router.ip" "Router/Gateway IP" "${GATEWAY_IP}"
prompt_if_unset_config "cloud.dns.ip" "DNS server IP (dnsmasq machine)" "${SUBNET_PREFIX}.50"
prompt_if_unset_config "cloud.dhcpRange" "DHCP range for dnsmasq" "${SUBNET_PREFIX}.100,${SUBNET_PREFIX}.200"
prompt_if_unset_config "cloud.dnsmasq.interface" "Network interface for dnsmasq" "eth0"
prompt_if_unset_config "cloud.dns.externalResolver" "External DNS resolver" "1.1.1.1"
# MetalLB IP address pool
prompt_if_unset_config "cluster.ipAddressPool" "MetalLB IP address pool" "${SUBNET_PREFIX}.80-${SUBNET_PREFIX}.89"
@@ -133,6 +127,10 @@ if [ -z "$current_lb_ip" ] || [ "$current_lb_ip" = "null" ]; then
print_info "Set load balancer IP to: ${lb_ip} (first IP in MetalLB pool)"
fi
# Talos version
prompt_if_unset_config "cluster.nodes.talos.version" "Talos version" "v1.10.4"
talos_version=$(wild-config "cluster.nodes.talos.version")
# Talos schematic ID
current_schematic_id=$(get_current_config "cluster.nodes.talos.schematicId")
if [ -z "$current_schematic_id" ] || [ "$current_schematic_id" = "null" ]; then
@@ -156,77 +154,40 @@ fi
cluster_name=$(get_current_config "cluster.name")
prompt_if_unset_config "cluster.externalDns.ownerId" "External DNS owner ID" "external-dns-${cluster_name}"
print_success "Cluster configuration completed"
echo ""
# =============================================================================
# Node Hardware Detection
# Node setup
# =============================================================================
if [ "${SKIP_HARDWARE}" = false ]; then
print_header "Node Hardware Detection"
print_info "This phase will help you register Talos nodes by discovering their hardware."
print_info "You'll need nodes booted in maintenance mode and accessible via IP."
echo ""
# Configure control plane network topology first
if [ -z "$(get_current_config "cluster.nodes.control.vip")" ]; then
print_header "Control Plane Network Configuration"
# Detect current network for suggestions
CURRENT_IP=$(ip route get 8.8.8.8 | awk '{print $7; exit}' 2>/dev/null || echo "192.168.1.100")
SUBNET_PREFIX=$(echo "${CURRENT_IP}" | cut -d. -f1-3)
print_info "Configure control plane nodes (you need at least 3 for HA):"
echo ""
prompt_if_unset_config "cluster.nodes.control.vip" "Control plane virtual IP" "${SUBNET_PREFIX}.90"
vip=$(wild-config "cluster.nodes.control.vip")
# Automatically configure the first three IPs after VIP for control plane nodes
vip_last_octet=$(echo "$vip" | cut -d. -f4)
vip_prefix=$(echo "$vip" | cut -d. -f1-3)
print_info "Configuring control plane nodes using consecutive IPs after VIP:"
for i in 1 2 3; do
node_ip="${vip_prefix}.$(( vip_last_octet + i ))"
print_info " Control plane node $i: $node_ip"
print_header "Control Plane Configuration"
# Initialize the node in cluster.nodes.active if not already present
if [ -z "$(get_current_config "cluster.nodes.active.\"${node_ip}\".control")" ]; then
wild-config-set "cluster.nodes.active.\"${node_ip}\".control" "true"
fi
done
print_success "Control plane network configuration completed"
echo ""
fi
# # Generate initial cluster configuration
# print_header "Cluster Configuration Generation"
# print_info "Generating base cluster configuration with talosctl gen config..."
# wild-cluster-config-generate
# Detect and register control plane nodes
print_header "Control Plane Node Registration"
# Get VIP to determine control plane IPs
vip=$(get_current_config "cluster.nodes.control.vip")
if [ -z "$vip" ]; then
print_error "VIP not configured. Run control plane network configuration first."
exit 1
fi
print_info "Configure control plane nodes (you need at least 3 for HA):"
echo ""
prompt_if_unset_config "cluster.nodes.control.vip" "Control plane virtual IP" "${SUBNET_PREFIX}.90"
vip=$(wild-config "cluster.nodes.control.vip")
# Automatically configure the first three IPs after VIP for control plane nodes
vip_last_octet=$(echo "$vip" | cut -d. -f4)
vip_prefix=$(echo "$vip" | cut -d. -f1-3)
# Detect and register control plane nodes
print_header "Control Plane Node Registration"
# Process each control plane node IP
for i in 1 2 3; do
TARGET_IP="${vip_prefix}.$(( vip_last_octet + i ))"
echo ""
print_info "Registering control plane node: $TARGET_IP"
# Initialize the node in cluster.nodes.active if not already present
if [ -z "$(get_current_config "cluster.nodes.active.\"${TARGET_IP}\".control")" ]; then
wild-config-set "cluster.nodes.active.\"${TARGET_IP}\".control" "true"
fi
# Check if node is already configured
existing_interface=$(get_current_config "cluster.nodes.active.\"${TARGET_IP}\".interface")
if [ -n "$existing_interface" ] && [ "$existing_interface" != "null" ]; then
@@ -234,25 +195,25 @@ if [ "${SKIP_HARDWARE}" = false ]; then
print_info " - Interface: $existing_interface"
print_info " - Disk: $(get_current_config "cluster.nodes.active.\"${TARGET_IP}\".disk")"
# Generate machine config for this node if necessary.
# Generate machine config patch for this node if necessary.
NODE_SETUP_DIR="${WC_HOME}/setup/cluster-nodes"
CONFIG_FILE="${NODE_SETUP_DIR}/final/${TARGET_IP}.yaml"
CONFIG_FILE="${NODE_SETUP_DIR}/patch/${TARGET_IP}.yaml"
if [ ! -f "$CONFIG_FILE" ]; then
print_info "Generating missing machine configuration for $TARGET_IP..."
if wild-cluster-node-machine-config-generate "$TARGET_IP"; then
print_success "Machine configuration generated for $TARGET_IP"
print_info "Generating missing machine configuration patch for $TARGET_IP..."
if wild-cluster-node-patch-generate "$TARGET_IP"; then
print_success "Machine configuration patch generated for $TARGET_IP"
else
print_warning "Failed to generate machine configuration for $TARGET_IP"
print_warning "Failed to generate machine configuration patch for $TARGET_IP"
fi
else
print_info " ✓ Machine config exists: $CONFIG_FILE"
print_info " ✓ Machine configuration patch exists: $CONFIG_FILE"
fi
continue
fi
read -p "Do you want to register control plane node $TARGET_IP now? (y/N): " -r register_node
read -p "Do you want to bring up control plane node $TARGET_IP now? (y/N): " -r register_node
if [[ ! $register_node =~ ^[Yy]$ ]]; then
print_info "Skipping node $TARGET_IP registration"
print_info "Skipping bringing up node $TARGET_IP registration"
continue
fi
@@ -319,7 +280,6 @@ if [ "${SKIP_HARDWARE}" = false ]; then
print_info "Updating configuration for $TARGET_IP..."
wild-config-set "cluster.nodes.active.\"${TARGET_IP}\".interface" "$INTERFACE"
wild-config-set "cluster.nodes.active.\"${TARGET_IP}\".disk" "$SELECTED_DISK"
wild-config-set "cluster.nodes.active.\"${TARGET_IP}\".control" "true"
# Copy current Talos version and schematic ID to this node
current_talos_version=$(get_current_config "cluster.nodes.talos.version")
@@ -330,34 +290,57 @@ if [ "${SKIP_HARDWARE}" = false ]; then
if [ -n "$current_schematic_id" ] && [ "$current_schematic_id" != "null" ]; then
wild-config-set "cluster.nodes.active.\"${TARGET_IP}\".schematicId" "$current_schematic_id"
fi
print_success "Node $TARGET_IP registered successfully"
# Generate machine config.
print_info "Generating machine configuration for $TARGET_IP..."
if wild-cluster-node-machine-config-generate "$TARGET_IP"; then
print_success "Machine configuration generated for $TARGET_IP"
# Ask if user wants to apply the configuration now
echo ""
read -p "Apply configuration to node $TARGET_IP now? (y/N): " -r apply_config
if [[ $apply_config =~ ^[Yy]$ ]]; then
if [ "$DETECTION_IP" != "$TARGET_IP" ]; then
# Node is in maintenance mode, use insecure flag
print_info "Applying configuration in insecure mode (maintenance mode)..."
wild-cluster-node-up "$TARGET_IP" --insecure
else
# Node is already configured, use secure mode
print_info "Applying configuration..."
wild-cluster-node-up "$TARGET_IP"
fi
echo ""
read -p "Bring node $TARGET_IP up now? (y/N): " -r apply_config
if [[ $apply_config =~ ^[Yy]$ ]]; then
if [ "$DETECTION_IP" != "$TARGET_IP" ]; then
# Node is in maintenance mode, use insecure flag
print_info "Applying configuration in insecure mode (maintenance mode)..."
wild-cluster-node-up "$TARGET_IP" --insecure
else
print_info "Configuration not applied. You can apply it later with:"
print_info " wild-cluster-node-up $TARGET_IP --insecure"
# Node is already configured, use secure mode
print_info "Applying configuration..."
wild-cluster-node-up "$TARGET_IP"
fi
# Bootstrap the cluster after the first node is up.
if [ "$i" -eq 1 ]; then
read -p "The cluster should be bootstrapped after the first control node is ready. Is it ready?: " -r is_ready
if [[ $is_ready =~ ^[Yy]$ ]]; then
print_info "Bootstrapping control plane node $TARGET_IP..."
talos config endpoint "$TARGET_IP"
# Attempt to bootstrap the cluster
if talosctl bootstrap --nodes "$TARGET_IP" 2>&1 | tee /tmp/bootstrap_output.log; then
print_success "Control plane node $TARGET_IP bootstrapped successfully!"
else
# Check if the error is because it's already bootstrapped
if grep -q "etcd data directory is not empty\|AlreadyExists" /tmp/bootstrap_output.log; then
print_info "Cluster is already bootstrapped on $TARGET_IP"
else
print_error "Failed to bootstrap control plane node $TARGET_IP"
print_info "Bootstrap output:"
cat /tmp/bootstrap_output.log
rm -f /tmp/bootstrap_output.log
continue
fi
fi
rm -f /tmp/bootstrap_output.log
talosctl config endpoint "$vip"
print_info "Talos endpoint set to control plane VIP: $vip"
talosctl kubeconfig "$vip"
print_success "Talos kubeconfig updated for control plane VIP: $vip"
fi
fi
else
print_warning "Failed to generate machine configuration for $TARGET_IP"
print_info "Configuration not applied. You can apply it later with:"
print_info " wild-cluster-node-up $TARGET_IP --insecure"
fi
fi
done