fix(installer): preserve Station Express recipe across host handoff

docs/get-started/dgx-station-preparation.mdx

@@ -68,7 +68,10 @@ Any other failed unit blocks preparation for administrator review.
 On the `--force-station-install` path, preparation also tolerates the InfiniBand communication manager (`ibacm`) and the realtime audio scheduler (`rtkit-daemon`) as failed units, because neither affects GPU or container capability; the required GPU, container, and preparation-critical service checks still apply.
 It does not install a host CUDA toolkit or Docker Compose.
 If any other existing prerequisite version differs, preparation stops instead of changing it automatically.
-After changing pinned packages, the installer exits with status `10`; reboot, sign in, and run the printed command, which pins the exact accepted NemoClaw commit before resuming express setup.
+Before host preparation begins, the installer stores the accepted Station Express recipe in owner-only local state.
+If preparation requires a reboot or a new login, run the printed command to restore the exact NemoClaw revision, agent, model, sandbox, policy tier, and gateway, dashboard, and vLLM ports without repeating the Express prompt.
+Recovery receipts created by earlier releases remain compatible.
+After changing pinned packages, the installer exits with status `10`; reboot, sign in, and run that printed command to resume Express setup.
 
 ## Validate Stock DGX OS
 

docs/get-started/quickstart.mdx

@@ -196,7 +196,9 @@ Use these details when your first-run path needs more control.
     DGX Spark uses managed vLLM with `qwen3.6-35b-a3b-nvfp4` by default.
     DGX Station express install explicitly selects `nemotron-3-ultra-550b-a55b` instead of the Station managed-vLLM profile default, `deepseek-v4-flash`, and discloses the approximately `352 GB` model download before confirmation.
     If the host does not meet the [Station prerequisites](prerequisites/dgx-station-preparation), the installer stops before host preparation.
-    Generic Ubuntu preparation can change pinned packages and then exits with status `10`; reboot, sign in, and run the printed exact-commit command to resume the accepted express recipe.
+    Before Station host preparation begins, the installer stores the accepted Express recipe in owner-only local state.
+    If preparation requires a reboot or a new login, run the printed command to restore the exact revision, agent, model, sandbox, policy tier, and gateway, dashboard, and vLLM ports without repeating the Express prompt.
+    Generic Ubuntu preparation can change pinned packages and then exits with status `10`; reboot, sign in, and run that printed command to resume.
     Stock DGX OS validation checks the factory stack in place without installing packages, restarting services, or rewriting the Docker runtime.
     One physical DGX OS `7.5.0` GB300 validation completed, but Station remains Deferred pending repeat clean-host qualification and CI coverage.
     To select DeepSeek V4 Flash while retaining the one-confirmation Station express flow, run `curl -fsSL https://www.nvidia.com/nemoclaw.sh | bash -s -- --station-deepseek`.

scripts/install.sh

@@ -197,7 +197,11 @@ resolve_nemoclaw_gateway_port() {
   local port="${NEMOCLAW_GATEWAY_PORT:-8080}"
   port="${port#"${port%%[![:space:]]*}"}"
   port="${port%"${port##*[![:space:]]}"}"
-  if [[ ! "$port" =~ ^[0-9]+$ ]] || [ "$port" -lt 1024 ] || [ "$port" -gt 65535 ]; then
+  if [[ ! "$port" =~ ^0*([0-9]{1,5})$ ]]; then
+    error "NEMOCLAW_GATEWAY_PORT must be an integer between 1024 and 65535."
+  fi
+  port="$((10#${BASH_REMATCH[1]}))"
+  if [ "$port" -lt 1024 ] || [ "$port" -gt 65535 ]; then
     error "NEMOCLAW_GATEWAY_PORT must be an integer between 1024 and 65535."
   fi
   if [ "$port" -ge 18789 ] && [ "$port" -le 18799 ]; then
@@ -3083,6 +3087,9 @@ _SELECTED_EXPRESS_PLATFORM=""
 _STATION_EXPRESS_RESUME_REVISION=""
 _STATION_EXPRESS_RESUME_LOADED=""
 _STATION_EXPRESS_RESUME_GENERATION=""
+_STATION_EXPRESS_RESUME_GATEWAY_PORT=""
+_STATION_EXPRESS_RESUME_DASHBOARD_PORT=""
+_STATION_EXPRESS_RESUME_VLLM_PORT=""
 
 normalize_station_vllm_model() {
   printf "%s" "${1:-}" | tr '[:upper:]' '[:lower:]' | sed 's/^[[:space:]]*//; s/[[:space:]]*$//'
@@ -3268,6 +3275,38 @@ validate_station_express_resume_generation() {
   [[ "${1:-}" =~ ^[0-9a-f]{32}$ ]]
 }
 
+validate_station_express_resume_port() {
+  local port="${1:-}"
+  [[ "$port" =~ ^[0-9]+$ ]] && [ "$port" -ge 1024 ] && [ "$port" -le 65535 ]
+}
+
+station_express_resume_port_value() {
+  local env_name="$1" fallback="$2" port
+  case "$env_name" in
+    NEMOCLAW_GATEWAY_PORT | NEMOCLAW_DASHBOARD_PORT | NEMOCLAW_VLLM_PORT) ;;
+    *) error "Unsupported DGX Station express resume port field: ${env_name}" ;;
+  esac
+  port="${!env_name:-$fallback}"
+  port="${port#"${port%%[![:space:]]*}"}"
+  port="${port%"${port##*[![:space:]]}"}"
+  validate_station_express_resume_port "$port" \
+    || error "${env_name} must be an integer between 1024 and 65535."
+  printf '%s' "$((10#$port))"
+}
+
+validate_station_express_resume_ports_distinct() {
+  local gateway_port="$1" dashboard_port="$2" vllm_port="$3"
+  gateway_port="$((10#$gateway_port))"
+  dashboard_port="$((10#$dashboard_port))"
+  vllm_port="$((10#$vllm_port))"
+  [[ "$gateway_port" != "$dashboard_port" ]] \
+    || error "NEMOCLAW_GATEWAY_PORT conflicts with NEMOCLAW_DASHBOARD_PORT (${gateway_port})."
+  [[ "$gateway_port" != "$vllm_port" ]] \
+    || error "NEMOCLAW_GATEWAY_PORT conflicts with NEMOCLAW_VLLM_PORT (${gateway_port})."
+  [[ "$dashboard_port" != "$vllm_port" ]] \
+    || error "NEMOCLAW_DASHBOARD_PORT conflicts with NEMOCLAW_VLLM_PORT (${dashboard_port})."
+}
+
 station_express_resume_generation() {
   local generation
   [[ -r /proc/sys/kernel/random/uuid ]] \
@@ -3330,7 +3369,9 @@ assert_station_express_resume_directory_safe() {
 
 load_station_express_resume() {
   local state_file revision_line model_line generation_line agent_line sandbox_line policy_tier_line
+  local gateway_port_line dashboard_port_line vllm_port_line
   local line_count saved_revision current_revision saved_agent saved_sandbox saved_policy_tier current_agent
+  local saved_gateway_port saved_dashboard_port saved_vllm_port current_gateway_port
   state_file="$(station_express_resume_file)" || return 1
   assert_nemoclaw_state_path_safe "$state_file"
   [[ -e "$state_file" || -L "$state_file" ]] || return 1
@@ -3342,6 +3383,9 @@ load_station_express_resume() {
   agent_line="$(sed -n '4p' "$state_file")"
   sandbox_line="$(sed -n '5p' "$state_file")"
   policy_tier_line="$(sed -n '6p' "$state_file")"
+  gateway_port_line="$(sed -n '7p' "$state_file")"
+  dashboard_port_line="$(sed -n '8p' "$state_file")"
+  vllm_port_line="$(sed -n '9p' "$state_file")"
   saved_revision="${revision_line#revision=}"
   NEMOCLAW_VLLM_MODEL="${model_line#model=}"
   _STATION_EXPRESS_RESUME_GENERATION="${generation_line#generation=}"
@@ -3353,44 +3397,98 @@ load_station_express_resume() {
     saved_agent="${agent_line#agent=}"
     saved_sandbox="${sandbox_line#sandbox=}"
     saved_policy_tier="${policy_tier_line#policy_tier=}"
+  elif [[ "$line_count" == "9" ]]; then
+    saved_agent="${agent_line#agent=}"
+    saved_sandbox="${sandbox_line#sandbox=}"
+    saved_policy_tier="${policy_tier_line#policy_tier=}"
+    saved_gateway_port="${gateway_port_line#gateway_port=}"
+    saved_dashboard_port="${dashboard_port_line#dashboard_port=}"
+    saved_vllm_port="${vllm_port_line#vllm_port=}"
   else
     error "DGX Station express resume state is invalid. Remove ${state_file} and rerun the installer."
   fi
   if [[ "$revision_line" != "revision=${saved_revision}" || "$model_line" != "model=${NEMOCLAW_VLLM_MODEL}" || "$generation_line" != "generation=${_STATION_EXPRESS_RESUME_GENERATION}" ]] \
-    || { [[ "$line_count" == "6" ]] && [[ "$agent_line" != "agent=${saved_agent}" || "$sandbox_line" != "sandbox=${saved_sandbox}" || "$policy_tier_line" != "policy_tier=${saved_policy_tier}" ]]; } \
+    || { [[ "$line_count" != "3" ]] && [[ "$agent_line" != "agent=${saved_agent}" || "$sandbox_line" != "sandbox=${saved_sandbox}" || "$policy_tier_line" != "policy_tier=${saved_policy_tier}" ]]; } \
+    || { [[ "$line_count" == "9" ]] && [[ "$gateway_port_line" != "gateway_port=${saved_gateway_port}" || "$dashboard_port_line" != "dashboard_port=${saved_dashboard_port}" || "$vllm_port_line" != "vllm_port=${saved_vllm_port}" ]]; } \
     || ! validate_station_express_resume_revision "$saved_revision" \
     || ! validate_station_express_resume_model "$NEMOCLAW_VLLM_MODEL" \
     || ! validate_station_express_resume_generation "$_STATION_EXPRESS_RESUME_GENERATION" \
     || ! validate_station_express_resume_agent "$saved_agent" \
     || ! validate_station_express_resume_sandbox "$saved_sandbox" \
-    || ! validate_station_express_resume_policy_tier "$saved_policy_tier"; then
+    || ! validate_station_express_resume_policy_tier "$saved_policy_tier" \
+    || { [[ "$line_count" == "9" ]] && { ! validate_station_express_resume_port "$saved_gateway_port" || ! validate_station_express_resume_port "$saved_dashboard_port" || ! validate_station_express_resume_port "$saved_vllm_port"; }; }; then
     error "DGX Station express resume state is invalid. Remove ${state_file} and rerun the installer."
   fi
+  current_gateway_port="$(resolve_nemoclaw_gateway_port)"
+  if [[ "$line_count" != "9" ]]; then
+    saved_gateway_port="$current_gateway_port"
+    saved_dashboard_port="$(station_express_resume_port_value NEMOCLAW_DASHBOARD_PORT 18789)"
+    saved_vllm_port="$(station_express_resume_port_value NEMOCLAW_VLLM_PORT 8000)"
+  fi
+  validate_station_express_resume_ports_distinct \
+    "$saved_gateway_port" "$saved_dashboard_port" "$saved_vllm_port"
+  _STATION_EXPRESS_RESUME_REVISION="$saved_revision"
+  _STATION_EXPRESS_RESUME_AGENT="$saved_agent"
+  _STATION_EXPRESS_RESUME_SANDBOX="$saved_sandbox"
+  _STATION_EXPRESS_RESUME_POLICY_TIER="$saved_policy_tier"
+  _STATION_EXPRESS_RESUME_GATEWAY_PORT="$saved_gateway_port"
+  _STATION_EXPRESS_RESUME_DASHBOARD_PORT="$saved_dashboard_port"
+  _STATION_EXPRESS_RESUME_VLLM_PORT="$saved_vllm_port"
   current_revision="$(station_installer_revision)"
   if [[ "$current_revision" != "$saved_revision" ]]; then
-    error "DGX Station express resume requires NemoClaw revision ${saved_revision}, but this installer is ${current_revision}. Rerun with: curl -fsSL https://www.nvidia.com/nemoclaw.sh | NEMOCLAW_INSTALL_TAG=${saved_revision} bash"
+    error "DGX Station Express resume requires NemoClaw revision ${saved_revision}, but this installer is ${current_revision}. Rerun with: $(station_express_resume_command)"
   fi
   current_agent="${NEMOCLAW_AGENT:-openclaw}"
   if [[ "$current_agent" != "$saved_agent" ]]; then
     error "DGX Station express resume requires NEMOCLAW_AGENT=${saved_agent}. Rerun the exact command printed after host preparation."
   fi
+  if [[ "$line_count" == "9" ]]; then
+    if [[ "$current_gateway_port" != "$saved_gateway_port" ]]; then
+      error "DGX Station express resume requires NEMOCLAW_GATEWAY_PORT=${saved_gateway_port}. Rerun the exact command printed after host preparation."
+    fi
+    if [[ -n "${NEMOCLAW_DASHBOARD_PORT:-}" ]] \
+      && [[ "$(station_express_resume_port_value NEMOCLAW_DASHBOARD_PORT 18789)" != "$saved_dashboard_port" ]]; then
+      error "DGX Station express resume requires NEMOCLAW_DASHBOARD_PORT=${saved_dashboard_port}. Rerun the exact command printed after host preparation."
+    fi
+    if [[ -n "${NEMOCLAW_VLLM_PORT:-}" ]] \
+      && [[ "$(station_express_resume_port_value NEMOCLAW_VLLM_PORT 8000)" != "$saved_vllm_port" ]]; then
+      error "DGX Station express resume requires NEMOCLAW_VLLM_PORT=${saved_vllm_port}. Rerun the exact command printed after host preparation."
+    fi
+  fi
   NEMOCLAW_SANDBOX_NAME="$saved_sandbox"
   NEMOCLAW_POLICY_TIER="$saved_policy_tier"
+  NEMOCLAW_GATEWAY_PORT="$saved_gateway_port"
+  NEMOCLAW_DASHBOARD_PORT="$saved_dashboard_port"
+  NEMOCLAW_VLLM_PORT="$saved_vllm_port"
   _STATION_EXPRESS_RESUME_LOADED=1
   export NEMOCLAW_VLLM_MODEL
   export NEMOCLAW_SANDBOX_NAME
   export NEMOCLAW_POLICY_TIER
+  export NEMOCLAW_GATEWAY_PORT
+  export NEMOCLAW_DASHBOARD_PORT
+  export NEMOCLAW_VLLM_PORT
   export NEMOCLAW_STATION_EXPRESS_RECEIPT_GENERATION="$_STATION_EXPRESS_RESUME_GENERATION"
 }
 
 save_station_express_resume() {
   local state_file state_dir temp_file revision generation
   local model="${NEMOCLAW_VLLM_MODEL:-}" agent="${NEMOCLAW_AGENT:-openclaw}"
   local sandbox="${NEMOCLAW_SANDBOX_NAME:-my-assistant}" policy_tier="${NEMOCLAW_POLICY_TIER:-balanced}"
+  local gateway_port dashboard_port vllm_port
   validate_station_express_resume_model "$model" || error "Cannot save an invalid DGX Station express model selector."
   validate_station_express_resume_agent "$agent" || error "Cannot save an invalid DGX Station express agent."
   validate_station_express_resume_sandbox "$sandbox" || error "Cannot save an invalid DGX Station express sandbox name."
   validate_station_express_resume_policy_tier "$policy_tier" || error "Cannot save an invalid DGX Station express policy tier."
+  gateway_port="$(resolve_nemoclaw_gateway_port)"
+  dashboard_port="$(station_express_resume_port_value NEMOCLAW_DASHBOARD_PORT 18789)"
+  vllm_port="$(station_express_resume_port_value NEMOCLAW_VLLM_PORT 8000)"
+  validate_station_express_resume_ports_distinct "$gateway_port" "$dashboard_port" "$vllm_port"
+  NEMOCLAW_GATEWAY_PORT="$gateway_port"
+  NEMOCLAW_DASHBOARD_PORT="$dashboard_port"
+  NEMOCLAW_VLLM_PORT="$vllm_port"
+  export NEMOCLAW_GATEWAY_PORT
+  export NEMOCLAW_DASHBOARD_PORT
+  export NEMOCLAW_VLLM_PORT
   revision="$(station_installer_revision)"
   state_file="$(station_express_resume_file)" || error "Could not resolve NemoClaw state for DGX Station express resume."
   state_dir="$(ensure_nemoclaw_state_dir)" || error "Could not prepare NemoClaw state for DGX Station express resume."
@@ -3404,8 +3502,9 @@ save_station_express_resume() {
     rm -f "$temp_file"
     error "Could not secure DGX Station express resume state under ${state_dir}."
   }
-  if ! printf 'revision=%s\nmodel=%s\ngeneration=%s\nagent=%s\nsandbox=%s\npolicy_tier=%s\n' \
-    "$revision" "$model" "$generation" "$agent" "$sandbox" "$policy_tier" >"$temp_file"; then
+  if ! printf 'revision=%s\nmodel=%s\ngeneration=%s\nagent=%s\nsandbox=%s\npolicy_tier=%s\ngateway_port=%s\ndashboard_port=%s\nvllm_port=%s\n' \
+    "$revision" "$model" "$generation" "$agent" "$sandbox" "$policy_tier" \
+    "$gateway_port" "$dashboard_port" "$vllm_port" >"$temp_file"; then
     rm -f "$temp_file"
     error "Could not write DGX Station express resume state under ${state_dir}."
   fi
@@ -3419,12 +3518,17 @@ save_station_express_resume() {
   _STATION_EXPRESS_RESUME_AGENT="$agent"
   _STATION_EXPRESS_RESUME_SANDBOX="$sandbox"
   _STATION_EXPRESS_RESUME_POLICY_TIER="$policy_tier"
+  _STATION_EXPRESS_RESUME_GATEWAY_PORT="$gateway_port"
+  _STATION_EXPRESS_RESUME_DASHBOARD_PORT="$dashboard_port"
+  _STATION_EXPRESS_RESUME_VLLM_PORT="$vllm_port"
 }
 
 station_express_resume_command() {
-  printf 'curl -fsSL https://www.nvidia.com/nemoclaw.sh | NEMOCLAW_INSTALL_TAG=%s NEMOCLAW_AGENT=%s NEMOCLAW_SANDBOX_NAME=%s NEMOCLAW_POLICY_TIER=%s bash' \
+  printf 'curl -fsSL https://www.nvidia.com/nemoclaw.sh | NEMOCLAW_INSTALL_TAG=%s NEMOCLAW_AGENT=%s NEMOCLAW_SANDBOX_NAME=%s NEMOCLAW_POLICY_TIER=%s NEMOCLAW_GATEWAY_PORT=%s NEMOCLAW_DASHBOARD_PORT=%s NEMOCLAW_VLLM_PORT=%s bash' \
     "$_STATION_EXPRESS_RESUME_REVISION" "$_STATION_EXPRESS_RESUME_AGENT" \
-    "$_STATION_EXPRESS_RESUME_SANDBOX" "$_STATION_EXPRESS_RESUME_POLICY_TIER"
+    "$_STATION_EXPRESS_RESUME_SANDBOX" "$_STATION_EXPRESS_RESUME_POLICY_TIER" \
+    "$_STATION_EXPRESS_RESUME_GATEWAY_PORT" "$_STATION_EXPRESS_RESUME_DASHBOARD_PORT" \
+    "$_STATION_EXPRESS_RESUME_VLLM_PORT"
   if [ "${FORCE_STATION_INSTALL:-}" = "1" ]; then
     printf ' -s -- --force-station-install'
   fi
@@ -3544,6 +3648,11 @@ filter_station_host_preparation_output() {
 ensure_station_express_host() {
   [[ "${_SELECTED_EXPRESS_PLATFORM:-}" == "DGX Station" ]] || return 0
 
+  # Publish the accepted, secret-free recipe before the helper can mutate the
+  # host. The same receipt then binds reboot/login continuation and onboarding
+  # recovery to this exact Express attempt.
+  save_station_express_resume
+
   local release_state
   release_state="$(classify_dgx_station_release)"
   case "$release_state" in
@@ -3569,17 +3678,15 @@ ensure_station_express_host() {
       ok "DGX Station host prerequisites are ready"
       ;;
     10)
-      save_station_express_resume
       warn "DGX Station host prerequisites were installed and require a reboot."
       info "Run: sudo reboot"
-      info "After signing in again, rerun the accepted revision:"
+      info "After signing in again, rerun the accepted Station Express recipe:"
       info "$(station_express_resume_command)"
       exit 10
       ;;
     11)
-      save_station_express_resume
       warn "Docker access was granted and requires a new login session. A reboot is not required."
-      info "After signing in again, rerun the accepted revision:"
+      info "After signing in again, rerun the accepted Station Express recipe:"
       info "$(station_express_resume_command)"
       exit 11
       ;;

src/lib/onboard/station-express-resume.test.ts

@@ -59,6 +59,16 @@ function currentReceiptText(
   return `${receiptText().trimEnd()}\nagent=${agent}\nsandbox=${sandbox}\npolicy_tier=${policyTier}\n`;
 }
 
+function portReceiptText(
+  overrides: Partial<{ gatewayPort: string; dashboardPort: string; vllmPort: string }> = {},
+): string {
+  const { gatewayPort = "18081", dashboardPort = "18790", vllmPort = "18000" } = overrides;
+  return (
+    `${currentReceiptText().trimEnd()}\n` +
+    `gateway_port=${gatewayPort}\ndashboard_port=${dashboardPort}\nvllm_port=${vllmPort}\n`
+  );
+}
+
 function retirementClaims(home: string): string[] {
   const stateDir = path.join(home, ".nemoclaw");
   return fs
@@ -645,6 +655,47 @@ describe("DGX Station Express resume (#7048)", () => {
     }
   });
 
+  it("accepts and retires the port-bound installer receipt (#7203)", () => {
+    const home = fs.mkdtempSync(path.join(os.tmpdir(), "nemoclaw-station-port-receipt-"));
+    const stateDir = path.join(home, ".nemoclaw");
+    const receipt = path.join(stateDir, "station-express-resume");
+    fs.mkdirSync(stateDir, { mode: 0o700 });
+    fs.writeFileSync(receipt, portReceiptText(), { mode: 0o600 });
+
+    try {
+      expect(() =>
+        assertStationExpressInstallerResumeMatches(receiptGeneration, { HOME: home }),
+      ).not.toThrow();
+      retireStationExpressInstallerResume(receiptGeneration, { env: { HOME: home } });
+      expect(fs.existsSync(receipt)).toBe(false);
+    } finally {
+      fs.rmSync(home, { recursive: true, force: true });
+    }
+  });
+
+  it.each([
+    ["gateway", { gatewayPort: "invalid" }],
+    ["dashboard", { dashboardPort: "1023" }],
+    ["vLLM", { vllmPort: "65536" }],
+    ["duplicate", { dashboardPort: "18081" }],
+    ["numerically duplicate", { dashboardPort: "018081" }],
+  ])("rejects a port-bound installer receipt with an invalid %s port", (_field, overrides) => {
+    const home = fs.mkdtempSync(path.join(os.tmpdir(), "nemoclaw-station-port-invalid-"));
+    const stateDir = path.join(home, ".nemoclaw");
+    const receipt = path.join(stateDir, "station-express-resume");
+    fs.mkdirSync(stateDir, { mode: 0o700 });
+    fs.writeFileSync(receipt, portReceiptText(overrides), { mode: 0o600 });
+
+    try {
+      expect(() =>
+        assertStationExpressInstallerResumeMatches(receiptGeneration, { HOME: home }),
+      ).toThrow("installer resume state is malformed");
+      expect(fs.readFileSync(receipt, "utf8")).toBe(portReceiptText(overrides));
+    } finally {
+      fs.rmSync(home, { recursive: true, force: true });
+    }
+  });
+
   it.each([
     ["agent", { agent: "unknown-agent" }],
     ["sandbox", { sandbox: "Invalid Sandbox" }],