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

Cluster nodes by name instead of (mutable) ip.

Browse Source
This commit is contained in:
Paul Payne
2025-07-21 17:32:21 -07:00
parent 9d1ad5950b
commit c0b1d60e7b
7 changed files with 197 additions and 158 deletions
Download Patch File Download Diff File Expand all files Collapse all files

105
bin/wild-setup-cluster
View File

@@ -103,6 +103,10 @@ fi
prompt_if_unset_config "operator.email" "Operator email address"
# Configure hostname prefix for unique node names on LAN
prompt_if_unset_config "cluster.hostnamePrefix" "Hostname prefix (optional, e.g. 'test-' for unique names on LAN)" ""
HOSTNAME_PREFIX=$(wild-config "cluster.hostnamePrefix")
# Configure network settings
CURRENT_IP=$(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 $3; exit}' 2>/dev/null || echo "192.168.1.1")
@@ -176,33 +180,36 @@ if [ "${SKIP_HARDWARE}" = false ]; then
# Detect and register control plane nodes
print_header "Control Plane Node Registration"
# Process each control plane node IP
# Process each control plane node
for i in 1 2 3; do
NODE_NAME="${HOSTNAME_PREFIX}control-${i}"
TARGET_IP="${vip_prefix}.$(( vip_last_octet + i ))"
echo ""
print_info "Registering control plane node: $TARGET_IP"
print_info "Registering control plane node: $NODE_NAME (IP: $TARGET_IP)"
# Initialize the node in cluster.nodes.active if not already present
if [ -z "$(get_current_config "cluster.nodes.active.\"${TARGET_IP}\".control")" ]; then
wild-config-set "cluster.nodes.active.\"${TARGET_IP}\".control" "true"
if [ -z "$(get_current_config "cluster.nodes.active.\"${NODE_NAME}\".role")" ]; then
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".role" "controlplane"
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".targetIp" "$TARGET_IP"
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".currentIp" "$TARGET_IP"
fi
# Check if node is already configured
existing_interface=$(get_current_config "cluster.nodes.active.\"${TARGET_IP}\".interface")
existing_interface=$(get_current_config "cluster.nodes.active.\"${NODE_NAME}\".interface")
if [ -n "$existing_interface" ] && [ "$existing_interface" != "null" ]; then
print_success "Node $TARGET_IP already configured"
print_success "Node $NODE_NAME already configured"
print_info " - Interface: $existing_interface"
print_info " - Disk: $(get_current_config "cluster.nodes.active.\"${TARGET_IP}\".disk")"
print_info " - Disk: $(get_current_config "cluster.nodes.active.\"${NODE_NAME}\".disk")"
# Generate machine config patch for this node if necessary.
NODE_SETUP_DIR="${WC_HOME}/setup/cluster-nodes"
CONFIG_FILE="${NODE_SETUP_DIR}/patch/${TARGET_IP}.yaml"
CONFIG_FILE="${NODE_SETUP_DIR}/patch/${NODE_NAME}.yaml"
if [ ! -f "$CONFIG_FILE" ]; then
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"
print_info "Generating missing machine configuration patch for $NODE_NAME..."
if wild-cluster-node-patch-generate "$NODE_NAME"; then
print_success "Machine configuration patch generated for $NODE_NAME"
else
print_warning "Failed to generate machine configuration patch for $TARGET_IP"
print_warning "Failed to generate machine configuration patch for $NODE_NAME"
fi
else
print_info " ✓ Machine configuration patch exists: $CONFIG_FILE"
@@ -210,9 +217,9 @@ if [ "${SKIP_HARDWARE}" = false ]; then
continue
fi
read -p "Do you want to bring up control plane node $TARGET_IP now? (y/N): " -r register_node
read -p "Do you want to bring up control plane node $NODE_NAME ($TARGET_IP) now? (y/N): " -r register_node
if [[ ! $register_node =~ ^[Yy]$ ]]; then
print_info "Skipping bringing up node $TARGET_IP registration"
print_info "Skipping bringing up node $NODE_NAME registration"
continue
fi
@@ -231,7 +238,7 @@ if [ "${SKIP_HARDWARE}" = false ]; then
read -p "Enter maintenance IP for this node: " -r MAINTENANCE_IP
if [ -z "$MAINTENANCE_IP" ]; then
print_warning "Skipping node $TARGET_IP registration"
print_warning "Skipping node $NODE_NAME registration"
continue
fi
@@ -241,7 +248,7 @@ if [ "${SKIP_HARDWARE}" = false ]; then
DETECTION_IP="$MAINTENANCE_IP"
# Store maintenance IP for reference
wild-config-set "cluster.nodes.active.\"${TARGET_IP}\".maintenanceIp" "$MAINTENANCE_IP"
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".maintenanceIp" "$MAINTENANCE_IP"
print_success "Node detected at maintenance IP $MAINTENANCE_IP"
else
print_error "Failed to detect node at $MAINTENANCE_IP"
@@ -276,31 +283,31 @@ if [ "${SKIP_HARDWARE}" = false ]; then
fi
# Update config.yaml with hardware info.
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"
print_info "Updating configuration for $NODE_NAME..."
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".interface" "$INTERFACE"
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".disk" "$SELECTED_DISK"
# Copy current Talos version and schematic ID to this node
current_talos_version=$(get_current_config "cluster.nodes.talos.version")
current_schematic_id=$(get_current_config "cluster.nodes.talos.schematicId")
if [ -n "$current_talos_version" ] && [ "$current_talos_version" != "null" ]; then
wild-config-set "cluster.nodes.active.\"${TARGET_IP}\".version" "$current_talos_version"
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".version" "$current_talos_version"
fi
if [ -n "$current_schematic_id" ] && [ "$current_schematic_id" != "null" ]; then
wild-config-set "cluster.nodes.active.\"${TARGET_IP}\".schematicId" "$current_schematic_id"
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".schematicId" "$current_schematic_id"
fi
echo ""
read -p "Bring node $TARGET_IP up now? (y/N): " -r apply_config
read -p "Bring node $NODE_NAME ($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 "$TARGET_IP" --insecure
wild-cluster-node-up "$NODE_NAME" --insecure
else
# Node is already configured, use secure mode
print_info "Applying configuration..."
wild-cluster-node-up "$TARGET_IP"
wild-cluster-node-up "$NODE_NAME"
fi
# Bootstrap the cluster after the first node is up.
@@ -337,7 +344,7 @@ if [ "${SKIP_HARDWARE}" = false ]; then
else
print_info "Configuration not applied. You can apply it later with:"
print_info " wild-cluster-node-up $TARGET_IP --insecure"
print_info " wild-cluster-node-up $NODE_NAME --insecure"
fi
fi
@@ -346,28 +353,30 @@ if [ "${SKIP_HARDWARE}" = false ]; then
# Register worker nodes
echo ""
print_info "Configure worker nodes (optional):"
WORKER_COUNT=1
while true; do
echo ""
read -p "Do you want to register a worker node? (y/N): " -r register_worker
if [[ $register_worker =~ ^[Yy]$ ]]; then
read -p "Enter maintenance IP for worker node: " -r WORKER_IP
NODE_NAME="${HOSTNAME_PREFIX}worker-${WORKER_COUNT}"
read -p "Enter current IP for worker node $NODE_NAME: " -r WORKER_IP
if [ -z "$WORKER_IP" ]; then
print_warning "No IP provided, skipping worker node"
continue
fi
print_info "Running wild-node-detect for worker node $WORKER_IP..."
print_info "Running wild-node-detect for worker node $NODE_NAME ($WORKER_IP)..."
# Run detection and capture both output and stderr for debugging
DETECTION_OUTPUT=$(mktemp)
DETECTION_ERROR=$(mktemp)
if wild-node-detect "$WORKER_IP" >"$DETECTION_OUTPUT" 2>"$DETECTION_ERROR"; then
WORKER_INFO=$(cat "$DETECTION_OUTPUT")
print_success "Worker node detected at IP $WORKER_IP"
print_success "Worker node $NODE_NAME detected at IP $WORKER_IP"
rm -f "$DETECTION_OUTPUT" "$DETECTION_ERROR"
else
print_error "Failed to detect hardware for worker node $WORKER_IP"
print_error "Failed to detect hardware for worker node $NODE_NAME ($WORKER_IP)"
print_info "Detection error output:"
cat "$DETECTION_ERROR" >&2
print_info "Make sure the node is running in maintenance mode and accessible"
@@ -381,7 +390,7 @@ if [ "${SKIP_HARDWARE}" = false ]; then
SELECTED_DISK=$(echo "$WORKER_INFO" | jq -r '.selected_disk')
AVAILABLE_DISKS=$(echo "$WORKER_INFO" | jq -r '.disks | join(", ")')
print_success "Hardware detected for worker node $WORKER_IP:"
print_success "Hardware detected for worker node $NODE_NAME:"
print_info " - Interface: $INTERFACE"
print_info " - Available disks: $AVAILABLE_DISKS"
print_info " - Selected disk: $SELECTED_DISK"
@@ -402,51 +411,55 @@ if [ "${SKIP_HARDWARE}" = false ]; then
fi
# Update config.yaml with worker hardware info
print_info "Updating config.yaml for worker node $WORKER_IP..."
print_info "Updating config.yaml for worker node $NODE_NAME..."
# Store under unified cluster.nodes.active.<ip-address>
wild-config-set "cluster.nodes.active.\"${WORKER_IP}\".interface" "$INTERFACE"
wild-config-set "cluster.nodes.active.\"${WORKER_IP}\".disk" "$SELECTED_DISK"
wild-config-set "cluster.nodes.active.\"${WORKER_IP}\".control" "false"
# Store under unified cluster.nodes.active.<node-name>
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".role" "worker"
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".currentIp" "$WORKER_IP"
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".interface" "$INTERFACE"
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".disk" "$SELECTED_DISK"
# Copy current Talos version and schematic ID to this node
current_talos_version=$(get_current_config "cluster.nodes.talos.version")
current_schematic_id=$(get_current_config "cluster.nodes.talos.schematicId")
if [ -n "$current_talos_version" ] && [ "$current_talos_version" != "null" ]; then
wild-config-set "cluster.nodes.active.\"${WORKER_IP}\".version" "$current_talos_version"
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".version" "$current_talos_version"
fi
if [ -n "$current_schematic_id" ] && [ "$current_schematic_id" != "null" ]; then
wild-config-set "cluster.nodes.active.\"${WORKER_IP}\".schematicId" "$current_schematic_id"
wild-config-set "cluster.nodes.active.\"${NODE_NAME}\".schematicId" "$current_schematic_id"
fi
print_success "Worker node $WORKER_IP registered successfully:"
print_success "Worker node $NODE_NAME registered successfully:"
print_info " - Name: $NODE_NAME"
print_info " - IP: $WORKER_IP"
print_info " - Interface: $INTERFACE"
print_info " - Disk: $SELECTED_DISK"
# Generate machine config immediately
print_info "Generating machine configuration for $WORKER_IP..."
if wild-cluster-node-patch-generate "$WORKER_IP"; then
print_success "Machine configuration generated for $WORKER_IP"
print_info "Generating machine configuration for $NODE_NAME..."
if wild-cluster-node-patch-generate "$NODE_NAME"; then
print_success "Machine configuration generated for $NODE_NAME"
# Ask if user wants to apply the configuration now
echo ""
read -p "Apply configuration to worker node $WORKER_IP now? (y/N): " -r apply_config
read -p "Apply configuration to worker node $NODE_NAME now? (y/N): " -r apply_config
if [[ $apply_config =~ ^[Yy]$ ]]; then
# Worker nodes are typically in maintenance mode during setup
print_info "Applying configuration in insecure mode (maintenance mode)..."
wild-cluster-node-up "$WORKER_IP" --insecure
wild-cluster-node-up "$NODE_NAME" --insecure
else
print_info "Configuration not applied. You can apply it later with:"
print_info " wild-cluster-node-up $WORKER_IP --insecure"
print_info " wild-cluster-node-up $NODE_NAME --insecure"
fi
else
print_warning "Failed to generate machine configuration for $WORKER_IP"
print_warning "Failed to generate machine configuration for $NODE_NAME"
fi
else
print_error "Failed to detect hardware for worker node $WORKER_IP"
print_error "Failed to detect hardware for worker node $NODE_NAME"
continue
fi
WORKER_COUNT=$((WORKER_COUNT + 1))
else
break
fi