hecate: auto-bootstrap kubeconfig for peer fallback

scripts/install.sh

@@ -72,6 +72,98 @@ read_hecate_role() {
   echo "${role}"
 }
 
+migration_yaml_lookup() {
+  local key="$1"
+  awk -F': *' -v k="${key}" '$1 == k {print $2; exit}' "${CONF_DIR}/hecate.yaml" 2>/dev/null || true
+}
+
+first_control_plane_name() {
+  awk '
+    /^control_planes:[[:space:]]*$/ {in_list=1; next}
+    in_list && /^[[:space:]]*-[[:space:]]*/ {gsub(/^[[:space:]]*-[[:space:]]*/, "", $0); print $0; exit}
+    in_list && /^[^[:space:]]/ {in_list=0}
+  ' "${CONF_DIR}/hecate.yaml" 2>/dev/null || true
+}
+
+lookup_node_host() {
+  local node="$1"
+  awk -F': *' -v n="${node}" '$1 == "  " n {print $2; exit}' "${CONF_DIR}/hecate.yaml" 2>/dev/null || true
+}
+
+ensure_hecate_kubeconfig() {
+  local kubeconfig_path
+  kubeconfig_path="$(migration_yaml_lookup "kubeconfig")"
+  if [[ -z "${kubeconfig_path}" ]]; then
+    kubeconfig_path="/etc/hecate/kubeconfig"
+  fi
+  install -d -m 0750 "$(dirname "${kubeconfig_path}")"
+
+  if [[ -s "${kubeconfig_path}" ]] && KUBECONFIG="${kubeconfig_path}" kubectl version --request-timeout=5s >/dev/null 2>&1; then
+    return 0
+  fi
+
+  if [[ -r /etc/rancher/k3s/k3s.yaml ]]; then
+    install -m 0600 /etc/rancher/k3s/k3s.yaml "${kubeconfig_path}"
+    echo "[install] refreshed kubeconfig from local /etc/rancher/k3s/k3s.yaml"
+    if KUBECONFIG="${kubeconfig_path}" kubectl version --request-timeout=5s >/dev/null 2>&1; then
+      return 0
+    fi
+  fi
+
+  local cp_name cp_host ssh_user ssh_port ssh_cfg ssh_key
+  cp_name="$(first_control_plane_name)"
+  if [[ -z "${cp_name}" ]]; then
+    echo "[install] warning: cannot infer control plane name; kubeconfig bootstrap skipped"
+    return 0
+  fi
+  cp_host="$(lookup_node_host "${cp_name}")"
+  if [[ -z "${cp_host}" ]]; then
+    cp_host="${cp_name}"
+  fi
+  ssh_user="$(migration_yaml_lookup "ssh_user")"
+  ssh_port="$(migration_yaml_lookup "ssh_port")"
+  ssh_cfg="$(migration_yaml_lookup "ssh_config_file")"
+  ssh_key="$(migration_yaml_lookup "ssh_identity_file")"
+  if [[ -z "${ssh_port}" ]]; then
+    ssh_port="2277"
+  fi
+
+  local target
+  target="${cp_host}"
+  if [[ -n "${ssh_user}" ]]; then
+    target="${ssh_user}@${cp_host}"
+  fi
+  local ssh_args=(
+    -o BatchMode=yes
+    -o ConnectTimeout=8
+    -o StrictHostKeyChecking=accept-new
+  )
+  if [[ -n "${ssh_cfg}" && -f "${ssh_cfg}" ]]; then
+    ssh_args+=(-F "${ssh_cfg}")
+  fi
+  if [[ -n "${ssh_key}" && -f "${ssh_key}" ]]; then
+    ssh_args+=(-i "${ssh_key}")
+  fi
+  if [[ -n "${ssh_port}" ]]; then
+    ssh_args+=(-p "${ssh_port}")
+  fi
+
+  local remote_cfg
+  if remote_cfg="$(ssh "${ssh_args[@]}" "${target}" "sudo cat /etc/rancher/k3s/k3s.yaml" 2>/dev/null)"; then
+    printf '%s\n' "${remote_cfg}" > "${kubeconfig_path}"
+    sed -Ei "s#server:[[:space:]]*https://127\\.0\\.0\\.1:6443#server: https://${cp_host}:6443#g" "${kubeconfig_path}" || true
+    chmod 0600 "${kubeconfig_path}"
+    echo "[install] bootstrapped kubeconfig from control plane ${cp_name} (${cp_host})"
+    if KUBECONFIG="${kubeconfig_path}" kubectl version --request-timeout=5s >/dev/null 2>&1; then
+      return 0
+    fi
+  else
+    echo "[install] warning: failed to fetch kubeconfig from ${cp_name} (${cp_host})"
+  fi
+
+  echo "[install] warning: kubeconfig at ${kubeconfig_path} is still not validated; local startup fallback may fail"
+}
+
 migrate_hecate_config() {
   if [[ ! -f "${CONF_DIR}/hecate.yaml" ]]; then
     return 0
@@ -353,6 +445,7 @@ else
   echo "[install] keeping existing config at ${CONF_DIR}/hecate.yaml"
 fi
 migrate_hecate_config
+ensure_hecate_kubeconfig
 
 echo "[install] installing systemd units"
 install -m 0644 deploy/systemd/hecate.service "${SYSTEMD_DIR}/hecate.service"