style(vm): fix indentation inside create_vm() function body

- vm-core.func: add rm -f WORK_FILE in all recovery paths (RETRY,
  SKIP_CUSTOMIZE, RETRY_SETTINGS, NEW_VMID) to prevent multi-GB temp
  file leaks in /tmp on failed retry
- debian-13-vm.sh: same WORK_FILE cleanup in inline recovery path
- debian-13-vm.sh: fix misleading 'will configure on first boot' msg
  when skipping virt-customize (no first-boot mechanism exists)

docs/vm-smart-recovery-spec.md

@@ -0,0 +1,301 @@
+# VM Smart Recovery — Arbeitsanweisung
+
+**Branch:** `feature/vm-smart-recovery` (basiert auf `main`)  
+**Verwandt:** `feature/smart-error-recovery` (LXC, PR #11221)  
+**Erstellt:** 2026-02-16
+
+---
+
+## 1. Ausgangslage
+
+### Architektur-Vergleich LXC vs. VM
+
+| Aspekt | LXC (fertig in PR #11221) | VM (offen) |
+|---|---|---|
+| Shared Code | `misc/build.func` (5577 Zeilen) | `misc/vm-core.func` (627 Zeilen) — **nur von `docker-vm.sh` genutzt** |
+| Anzahl Scripts | ~170 | 15 |
+| Architektur | Alle nutzen `build_container()` | **2 Generationen** (s.u.) |
+| Software-Install | `pct exec` → Install-Script im Container | Variiert: `virt-customize`, Cloud-Init, `qm sendkey`, oder gar nichts |
+| Telemetrie | `post_to_api()` + `post_update_to_api()` | Identisch — alle sourcen `misc/api.func` |
+| Error Handling | Zentral in `build.func` Traps | Jedes Script hat eigenen `error_handler()` |
+| Recovery | Smart-Menü mit 6 dynamischen Optionen | **Keine** — bei Fehler wird VM sofort zerstört (`cleanup_vmid`) |
+
+### Zwei Generationen von VM-Scripts
+
+**Generation 1 — Legacy (monolithisch):** `haos-vm.sh`, `debian-vm.sh`, `openwrt-vm.sh` und 11 weitere.  
+- Selbstständige 500–700-Zeilen-Scripts
+- Definieren **alle** Utility-Funktionen inline (Colors, Icons, `msg_info`/`msg_ok`, `error_handler`, `cleanup`, etc.)
+- Sourcen nur `misc/api.func` für Telemetrie
+
+**Generation 2 — Modern (modular):** Ausschließlich `docker-vm.sh`.  
+- Sourced drei Shared Libraries:
+  - `misc/api.func` — Telemetrie
+  - `misc/vm-core.func` — Shared Utilities (627 Zeilen)
+  - `misc/cloud-init.func` — Cloud-Init Konfiguration (709 Zeilen)
+- Ruft `load_functions` aus `vm-core.func` auf
+
+### Telemetrie-Daten (Top VM-Failures)
+
+| Script | Anteil an VM-Failures |
+|---|---|
+| `docker-vm.sh` | 30.1 % |
+| `openwrt-vm.sh` | 25.9 % |
+| `debian-13-vm.sh` | 9.6 % |
+
+---
+
+## 2. Scope & Abgrenzung
+
+### In Scope
+
+- Smart Recovery für VM-Erstellungsfehler (Retry-Menü analog LXC)
+- Fehlererkennung: Download, Disk-Import, virt-customize, Ressourcen-Konflikte, Netzwerk
+- Exit-Code-Mapping (bereits in `api.func` vorhanden, wird geteilt)
+
+### Out of Scope (bewusst)
+
+- **Migration aller Legacy-Scripts auf `vm-core.func`** → eigenes Refactoring-Ticket
+- **In-VM-Repair** → VMs haben kein `pct exec`-Äquivalent
+- **`qm sendkey`-Recovery** (OpenWrt) → prinzipbedingt nicht retryable
+- **APT/DPKG-Repair innerhalb der VM** → kein Shell-Zugang während Install
+
+---
+
+## 3. Software-Installationsmethoden pro Script
+
+| Script | Methode | Beschreibung |
+|---|---|---|
+| `docker-vm.sh` | `virt-customize` | Offline Image-Manipulation (libguestfs) |
+| `docker-vm.sh` (Fallback) | systemd First-Boot-Service | Script läuft in VM beim ersten Boot |
+| `haos-vm.sh` | Keine | Pre-built Appliance (qcow2) |
+| `debian-vm.sh` / `debian-13-vm.sh` | Keine / Cloud-Init | Basis Cloud-Image |
+| `openwrt-vm.sh` | `qm sendkey` | Virtuelle Tastatur-Automation |
+| `opnsense-vm.sh` | `qm sendkey` + Bootstrap | Virtuelle Tastatur |
+| `ubuntu-*-vm.sh` | Cloud-Init | User konfiguriert vor Start |
+| `owncloud-vm.sh` | `virt-customize` | Wie docker-vm.sh |
+
+---
+
+## 4. Dateien & Änderungen
+
+### 4.1 `misc/vm-core.func` — Zentrale Recovery-Logik
+
+#### Neue Funktion: `vm_error_handler_with_recovery()`
+
+```
+Ablauf:
+├── Exit-Code erfassen ($? als ERSTES — kein ensure_log_on_host davor!)
+├── Fehlerklassifikation:
+│   ├── Download-Fehler (curl exit 6/7/22/28/35/52/56)
+│   ├── Disk-Import-Fehler (qm importdisk, pvesm alloc)
+│   ├── virt-customize-Fehler (libguestfs)
+│   ├── Ressourcen-Konflikt (VMID exists, Storage full)
+│   └── Netzwerk-Fehler (DNS, Timeout)
+├── Smart Recovery Menü:
+│   ├── [1] Retry (VM zerstören & neu erstellen)
+│   ├── [2] Retry mit anderen Einstellungen (RAM/CPU/Disk ändern)
+│   ├── [3] VM behalten (nicht zerstören, manuell debuggen)
+│   ├── [4] Abbrechen (VM zerstören, Exit)
+│   └── Dynamische Optionen je nach Fehlertyp:
+│       ├── Download-Fehler → "Cache löschen & neu downloaden"
+│       └── Ressourcen-Konflikt → "Andere VMID wählen"
+└── Bei Retry: cleanup_vmid() + create-Funktion erneut aufrufen
+```
+
+#### Neue Helper-Funktionen (Fehlererkennung):
+
+```bash
+is_download_error()      # curl exit codes + HTTP 404/500
+is_disk_import_error()   # qm importdisk stderr patterns
+is_virt_customize_err()  # libguestfs error patterns
+is_vmid_conflict()       # "already exists" in stderr
+is_storage_full()        # "not enough space" patterns
+```
+
+#### Log-Erfassung für VMs
+
+Anders als LXC (wo `/root/.install*.log` im Container liegt) müssen VM-Fehler direkt aus stderr der `qm`/`virt-customize` Befehle erfasst werden:
+
+```bash
+# Jeder kritische Befehl mit stderr-Capture:
+VM_ERROR_LOG="/tmp/vm-install-${VMID}.log"
+qm importdisk "$VMID" "$IMAGE" "$STORAGE" 2>> "$VM_ERROR_LOG"
+virt-customize -a "$IMAGE" --install docker.io 2>> "$VM_ERROR_LOG"
+```
+
+### 4.2 Retry-Wrapper-Architektur
+
+Da VMs kein zentrales `build_container()` haben, gibt es zwei Ansätze:
+
+#### Option A: Wrapper in `vm-core.func` (empfohlen für Gen-2 Scripts)
+
+```bash
+vm_create_with_recovery() {
+  local create_fn="$1"   # VM-spezifische Erstellungsfunktion
+  local max_retries=2
+  local attempt=0
+
+  while true; do
+    if "$create_fn"; then
+      return 0  # Erfolg
+    fi
+    ((attempt++))
+    if ((attempt >= max_retries)); then
+      # Max retries erreicht → nur noch "behalten" oder "abbrechen"
+    fi
+    vm_show_recovery_menu "$?" "$attempt"
+    # Menü-Auswahl verarbeiten...
+  done
+}
+```
+
+#### Option B: Inline-Recovery in Legacy-Scripts
+
+Für die 14 Legacy-Scripts (bis Migration auf `vm-core.func`):
+- Minimaler Patch: `error_handler()` um Recovery-Prompt erweitern
+- `cleanup_vmid` **nicht** sofort aufrufen, sondern erst nach User-Entscheidung
+
+**Empfehlung:** Zunächst **nur `docker-vm.sh`** (30.1 % der Failures) mit Option A umsetzen. Legacy-Scripts als Phase 2 nach Migration.
+
+### 4.3 `misc/api.func` — Keine Änderungen nötig
+
+Exit-Code-Mapping (`explain_exit_code()`) und `categorize_error()` sind bereits universal (LXC + VM). Nach Merge von PR #11221 stehen 70+ Exit-Codes zur Verfügung. Falls dieser Branch vorher fertig ist, können die Codes aus `feature/smart-error-recovery` cherry-picked werden.
+
+---
+
+## 5. Wichtige Unterschiede LXC vs. VM Recovery
+
+| LXC | VM |
+|---|---|
+| APT/DPKG In-Place-Repair im Container | **Nicht möglich** — kein Shell-Zugang während Install |
+| OOM-Retry mit x2 Ressourcen | **Funktioniert** — `qm set` kann RAM/CPU nachträglich ändern |
+| DNS-Override im Container (`/etc/resolv.conf`) | **Nicht anwendbar** — VM hat eigenes Netzwerk |
+| Container bleibt erhalten bei Repair | VM muss bei Retry **komplett neu erstellt** werden |
+| `build_container()` als zentrale Retry-Schleife | **Neue Wrapper-Funktion nötig** (`vm_create_with_recovery`) |
+| `pct exec` für In-Container-Zugriff | Kein Äquivalent (qemu-guest-agent nur wenn VM läuft + Agent installiert) |
+
+---
+
+## 6. Technische Fallstricke
+
+### 6.1 VMID-Cleanup vor Retry
+
+`cleanup_vmid` muss vollständig aufräumen:
+- `qm stop "$VMID" --skiplock` (falls Running)
+- `qm destroy "$VMID" --destroy-unreferenced-disks --purge`
+- Einige Scripts erzeugen zusätzliche Disks (`efidisk0`, `cloudinit`), die extra entfernt werden müssen
+
+### 6.2 Image-Caching
+
+`docker-vm.sh` cached Images in `/var/lib/vz/template/cache/`. Bei Download-Retry:
+- **Behalten**, wenn Download vollständig war (md5/sha-Check)
+- **Löschen**, wenn Corruption vermutet (curl-Fehler, xz-Validierung fehlgeschlagen)
+
+### 6.3 Cloud-Init-State
+
+Wenn Cloud-Init teilweise konfiguriert wurde, muss bei Retry der gesamte State zurückgesetzt werden:
+```bash
+qm set "$VMID" --delete cicustom
+qm set "$VMID" --delete ciuser
+qm set "$VMID" --delete cipassword
+```
+
+### 6.4 Legacy-Scripts (14 Stück)
+
+- Definieren `error_handler()` inline und sourcen nur `api.func`
+- Um dort Recovery einzubauen, entweder:
+  - **Jedes Script einzeln patchen** (hohes Risiko, viel Duplikat)
+  - **Erst Migration auf `vm-core.func`** (sauberer, aber größerer Scope)
+- **Empfehlung:** Migration priorisieren, Recovery danach trivial
+
+### 6.5 `virt-customize` Fallback
+
+`docker-vm.sh` hat bereits einen First-Boot-Fallback für Docker-Installation. Wenn `virt-customize` fehlschlägt:
+- Recovery sollte dies als **"soft failure"** behandeln
+- Aktiv den Fallback vorschlagen statt blindes Retry
+
+### 6.6 Kein `pct exec`-Äquivalent
+
+- Man kann **nicht "in die VM hinein reparieren"** wie bei LXC
+- `qm guest exec` existiert zwar (mit qemu-guest-agent), funktioniert aber nur wenn:
+  - Die VM läuft
+  - Der Guest Agent installiert ist
+  - Genau das ist typischerweise der Punkt, an dem der Install fehlschlägt
+
+---
+
+## 7. Implementierungsreihenfolge
+
+| Phase | Task | Dateien | Impact |
+|---|---|---|---|
+| **Phase 1** | `vm_error_handler_with_recovery()` Grundgerüst | `misc/vm-core.func` | Basis für alles |
+| **Phase 2** | `docker-vm.sh`: Recovery integrieren | `vm/docker-vm.sh` | 30.1 % der Failures |
+| **Phase 3** | Fehlererkennung (Download, Import, virt-customize) | `misc/vm-core.func` | Intelligente dynamische Menüoptionen |
+| **Phase 4** | `haos-vm.sh`: Recovery integrieren (Download-Retry) | `vm/haos-vm.sh` | Download-Corruption bereits teilweise vorhanden |
+| **Phase 5** | `debian-13-vm.sh` + `ubuntu-*-vm.sh` | `vm/debian-13-vm.sh`, etc. | Cloud-Image-Scripts |
+| **Phase 6** | `openwrt-vm.sh` (limitiert — nur Download/Import-Retry) | `vm/openwrt-vm.sh` | `sendkey`-Teil nicht retryable |
+
+---
+
+## 8. Test-Matrix
+
+| Szenario | Erwartetes Verhalten |
+|---|---|
+| Download-Fehler (curl 6/7/28) | Menü: "Retry Download" + "Cache löschen" |
+| Disk-Import-Fehler | Menü: "Retry" + "Anderen Storage wählen" |
+| VMID-Konflikt | Menü: "Andere VMID" + "Bestehende VM zerstören" |
+| virt-customize-Fehler (docker-vm) | Menü: "Retry" + "First-Boot-Fallback nutzen" |
+| Storage voll | Menü: "Anderen Storage wählen" + "Disk verkleinern" |
+| Netzwerk-Timeout | Menü: "Retry" + "Abbrechen" |
+| 2× Retry erreicht | Nur noch "VM behalten" oder "Abbrechen" |
+| User wählt "VM behalten" | VM nicht zerstören, manuellen Zugang erklären |
+
+---
+
+## 9. Branch-Workflow
+
+```bash
+# Neuen Branch erstellen (bereits geschehen):
+git checkout main
+git pull origin main
+git checkout -b feature/vm-smart-recovery
+
+# Arbeit in Phasen committen:
+# Phase 1: git commit -m "feat(vm): add vm_error_handler_with_recovery to vm-core.func"
+# Phase 2: git commit -m "feat(vm): integrate smart recovery into docker-vm.sh"
+# etc.
+
+# PR gegen main erstellen (NICHT gegen feature/smart-error-recovery)
+```
+
+### Abhängigkeit zu PR #11221
+
+Die `api.func`-Änderungen aus `feature/smart-error-recovery` (70+ Exit-Codes, `categorize_error()`) werden nach Merge von PR #11221 automatisch in `main` verfügbar sein.
+
+- Falls VM-Branch **nach** PR #11221 Merge gestartet wird → alles da
+- Falls VM-Branch **vorher** fertig ist → `api.func` Codes aus `feature/smart-error-recovery` cherry-picken
+
+---
+
+## 10. Referenz: Exit-0-Bug (nur LXC, gefixt)
+
+> Dieser Bug betrifft **nur LXC** (`misc/build.func`), nicht die VM-Scripts.
+
+**Root Cause:** Der ERR-Trap in `build.func` rief `ensure_log_on_host` vor `post_update_to_api` auf. Da `ensure_log_on_host` mit Exit 0 returned, wurde `$?` auf 0 zurückgesetzt → Telemetrie meldete "failed/0" statt dem echten Exit-Code (~15-20 Records/Tag).
+
+**Fix (PR #11221, Commit `2d7e707a0`):**
+```bash
+# Vorher (Bug):
+trap 'ensure_log_on_host; post_update_to_api "failed" "$?"' ERR
+
+# Nachher (Fix):
+trap '_ERR_CODE=$?; ensure_log_on_host; post_update_to_api "failed" "$_ERR_CODE"' ERR
+```
+
+**VM-Scripts nicht betroffen:** Diese erfassen `$?` korrekt als erste Zeile in `error_handler()`:
+```bash
+function error_handler() {
+  local exit_code="$?"  # Erste Zeile → korrekt
+  ...
+}
+```

misc/vm-core.func

@@ -624,3 +624,417 @@ EOF
   qm set "$VMID" -description "$DESCRIPTION" >/dev/null
 
 }
+
+# ==============================================================================
+# SECTION: VM SMART RECOVERY
+# ==============================================================================
+
+# Global error log for VM creation — captures stderr from critical commands
+VM_ERROR_LOG="${VM_ERROR_LOG:-/tmp/vm-install-$$.log}"
+VM_RECOVERY_ATTEMPT=${VM_RECOVERY_ATTEMPT:-0}
+VM_MAX_RETRIES=${VM_MAX_RETRIES:-2}
+
+# ------------------------------------------------------------------------------
+# vm_log_cmd()
+#
+# - Wraps a command to capture stderr into VM_ERROR_LOG
+# - Passes stdout through normally
+# - Returns the original exit code
+# Usage: vm_log_cmd qm importdisk "$VMID" "$IMAGE" "$STORAGE"
+# ------------------------------------------------------------------------------
+vm_log_cmd() {
+  "$@" 2>>"$VM_ERROR_LOG"
+}
+
+# ------------------------------------------------------------------------------
+# is_vm_download_error()
+#
+# - Detects download failures based on exit code and error log
+# - Checks curl exit codes (6, 7, 22, 28, 35, 52, 56) and HTTP error patterns
+# - Returns 0 (true) if download error detected, 1 otherwise
+# ------------------------------------------------------------------------------
+is_vm_download_error() {
+  local exit_code="${1:-0}"
+  local log_file="${2:-$VM_ERROR_LOG}"
+
+  # curl-specific exit codes indicating download issues
+  case "$exit_code" in
+  6 | 7 | 22 | 28 | 35 | 52 | 56) return 0 ;;
+  esac
+
+  # Check log for download-related patterns
+  if [[ -s "$log_file" ]]; then
+    if grep -qiE "curl.*failed|download.*failed|HTTP.*[45][0-9]{2}|Could not resolve|Connection refused|Connection timed out|SSL.*error" "$log_file" 2>/dev/null; then
+      return 0
+    fi
+  fi
+  return 1
+}
+
+# ------------------------------------------------------------------------------
+# is_vm_disk_import_error()
+#
+# - Detects disk import failures (qm importdisk / qm disk import)
+# - Checks for storage allocation and format conversion errors
+# - Returns 0 (true) if disk import error detected, 1 otherwise
+# ------------------------------------------------------------------------------
+is_vm_disk_import_error() {
+  local exit_code="${1:-0}"
+  local log_file="${2:-$VM_ERROR_LOG}"
+
+  if [[ -s "$log_file" ]]; then
+    if grep -qiE "importdisk.*failed|disk import.*error|storage.*allocation.*failed|qcow2.*error|raw.*error|pvesm.*alloc.*failed|unable to create|volume.*already exists" "$log_file" 2>/dev/null; then
+      return 0
+    fi
+  fi
+  return 1
+}
+
+# ------------------------------------------------------------------------------
+# is_vm_virt_customize_error()
+#
+# - Detects virt-customize / libguestfs failures
+# - Checks for guestfs, supermin, appliance boot errors
+# - Returns 0 (true) if virt-customize error detected, 1 otherwise
+# ------------------------------------------------------------------------------
+is_vm_virt_customize_error() {
+  local exit_code="${1:-0}"
+  local log_file="${2:-$VM_ERROR_LOG}"
+
+  if [[ -s "$log_file" ]]; then
+    if grep -qiE "virt-customize|libguestfs|guestfs|supermin|appliance.*boot|virt-.*failed|launch.*failed" "$log_file" 2>/dev/null; then
+      return 0
+    fi
+  fi
+  return 1
+}
+
+# ------------------------------------------------------------------------------
+# is_vm_vmid_conflict()
+#
+# - Detects VMID conflicts (VM already exists)
+# - Returns 0 (true) if conflict detected, 1 otherwise
+# ------------------------------------------------------------------------------
+is_vm_vmid_conflict() {
+  local exit_code="${1:-0}"
+  local log_file="${2:-$VM_ERROR_LOG}"
+
+  if [[ -s "$log_file" ]]; then
+    if grep -qiE "already exists|VM $VMID already|unable to create VM|VMID.*in use" "$log_file" 2>/dev/null; then
+      return 0
+    fi
+  fi
+  return 1
+}
+
+# ------------------------------------------------------------------------------
+# is_vm_storage_full()
+#
+# - Detects storage full / space exhaustion errors
+# - Returns 0 (true) if storage space issue detected, 1 otherwise
+# ------------------------------------------------------------------------------
+is_vm_storage_full() {
+  local exit_code="${1:-0}"
+  local log_file="${2:-$VM_ERROR_LOG}"
+
+  if [[ -s "$log_file" ]]; then
+    if grep -qiE "not enough space|no space left|storage.*full|disk quota|ENOSPC|insufficient.*space|thin pool.*full" "$log_file" 2>/dev/null; then
+      return 0
+    fi
+  fi
+  return 1
+}
+
+# ------------------------------------------------------------------------------
+# is_vm_network_error()
+#
+# - Detects general network/DNS errors beyond download failures
+# - Returns 0 (true) if network issue detected, 1 otherwise
+# ------------------------------------------------------------------------------
+is_vm_network_error() {
+  local exit_code="${1:-0}"
+  local log_file="${2:-$VM_ERROR_LOG}"
+
+  # Network-related curl/wget exit codes
+  case "$exit_code" in
+  6 | 7 | 28 | 52 | 56) return 0 ;;
+  esac
+
+  if [[ -s "$log_file" ]]; then
+    if grep -qiE "Name or service not known|Temporary failure in name resolution|Network is unreachable|No route to host|DNS.*failed|could not resolve" "$log_file" 2>/dev/null; then
+      return 0
+    fi
+  fi
+  return 1
+}
+
+# ------------------------------------------------------------------------------
+# vm_classify_error()
+#
+# - Classifies a VM creation error into a category
+# - Order matters: most specific checks first
+# - Returns category string via stdout
+# - Categories: vmid_conflict, storage_full, download, disk_import,
+#               virt_customize, network, unknown
+# ------------------------------------------------------------------------------
+vm_classify_error() {
+  local exit_code="${1:-0}"
+  local log_file="${2:-$VM_ERROR_LOG}"
+
+  if is_vm_vmid_conflict "$exit_code" "$log_file"; then
+    echo "vmid_conflict"
+  elif is_vm_storage_full "$exit_code" "$log_file"; then
+    echo "storage_full"
+  elif is_vm_download_error "$exit_code" "$log_file"; then
+    echo "download"
+  elif is_vm_disk_import_error "$exit_code" "$log_file"; then
+    echo "disk_import"
+  elif is_vm_virt_customize_error "$exit_code" "$log_file"; then
+    echo "virt_customize"
+  elif is_vm_network_error "$exit_code" "$log_file"; then
+    echo "network"
+  else
+    echo "unknown"
+  fi
+}
+
+# ------------------------------------------------------------------------------
+# vm_show_recovery_menu()
+#
+# - Displays a whiptail menu with recovery options after a VM creation failure
+# - Options are dynamically built based on error category
+# - Returns the selected option via stdout
+# - Arguments:
+#   $1: exit_code
+#   $2: error_category (from vm_classify_error)
+#   $3: current attempt number
+# ------------------------------------------------------------------------------
+vm_show_recovery_menu() {
+  local exit_code="${1:-1}"
+  local error_category="${2:-unknown}"
+  local attempt="${3:-1}"
+
+  local menu_items=()
+  local menu_height=12
+  local item_count=0
+
+  # --- Dynamic options based on error category ---
+
+  # Retry (always available unless max retries reached)
+  if ((attempt < VM_MAX_RETRIES)); then
+    case "$error_category" in
+    download)
+      menu_items+=("RETRY_DOWNLOAD" "🔄 Retry download (clear cache & re-download)" "ON")
+      ((item_count++))
+      ;;
+    disk_import)
+      menu_items+=("RETRY" "🔄 Retry VM creation" "ON")
+      ((item_count++))
+      ;;
+    virt_customize)
+      menu_items+=("RETRY" "🔄 Retry VM creation" "ON")
+      ((item_count++))
+      menu_items+=("SKIP_CUSTOMIZE" "⏭️  Skip virt-customize (use first-boot fallback)" "OFF")
+      ((item_count++))
+      ;;
+    network)
+      menu_items+=("RETRY" "🔄 Retry VM creation" "ON")
+      ((item_count++))
+      ;;
+    vmid_conflict)
+      menu_items+=("NEW_VMID" "🆔 Choose a different VM ID" "ON")
+      ((item_count++))
+      ;;
+    storage_full)
+      menu_items+=("RETRY_SETTINGS" "⚙️  Retry with different settings (storage/disk)" "ON")
+      ((item_count++))
+      ;;
+    *)
+      menu_items+=("RETRY" "🔄 Retry VM creation" "ON")
+      ((item_count++))
+      ;;
+    esac
+
+    # Retry with different resources (always offered)
+    menu_items+=("RETRY_SETTINGS" "⚙️  Retry with different settings (RAM/CPU/Disk)" "OFF")
+    ((item_count++))
+  fi
+
+  # Keep VM for debugging (always available)
+  menu_items+=("KEEP" "🔍 Keep partial VM for manual debugging" "OFF")
+  ((item_count++))
+
+  # Abort (always available)
+  menu_items+=("ABORT" "❌ Destroy VM and exit" "OFF")
+  ((item_count++))
+
+  menu_height=$((item_count + 10))
+
+  # Error info for title
+  local title="VM CREATION FAILED"
+  local body="Exit code: ${exit_code} | Category: ${error_category}\nAttempt: ${attempt}/${VM_MAX_RETRIES}\n\nChoose a recovery action:"
+
+  if ((attempt >= VM_MAX_RETRIES)); then
+    body="Exit code: ${exit_code} | Category: ${error_category}\n⚠️  Maximum retries (${VM_MAX_RETRIES}) reached.\n\nChoose an action:"
+  fi
+
+  local choice
+  choice=$(whiptail --backtitle "Proxmox VE Helper Scripts" --title "$title" \
+    --radiolist "$body" "$menu_height" 72 "$item_count" \
+    "${menu_items[@]}" 3>&1 1>&2 2>&3) || choice="ABORT"
+
+  echo "$choice"
+}
+
+# ------------------------------------------------------------------------------
+# vm_handle_recovery()
+#
+# - Main recovery handler called from error_handler or a wrapper
+# - Classifies the error, shows recovery menu, and executes the chosen action
+# - Arguments:
+#   $1: exit_code
+#   $2: line_number
+#   $3: failed_command
+#   $4: cleanup_fn     — function to call for VM cleanup (default: cleanup_vmid)
+#   $5: retry_fn       — function to re-invoke for full retry (required for retry)
+# - Uses global: VM_ERROR_LOG, VM_RECOVERY_ATTEMPT, VM_MAX_RETRIES, VMID
+# - Returns: 0 if retry was chosen (caller should re-run), 1 if abort/keep
+# ------------------------------------------------------------------------------
+vm_handle_recovery() {
+  local exit_code="${1:-1}"
+  local line_number="${2:-?}"
+  local failed_command="${3:-unknown}"
+  local cleanup_fn="${4:-cleanup_vmid}"
+  local retry_fn="${5:-}"
+
+  # Stop any running spinner
+  stop_spinner 2>/dev/null || true
+
+  # Classify the error
+  local error_category
+  error_category=$(vm_classify_error "$exit_code" "$VM_ERROR_LOG")
+
+  ((VM_RECOVERY_ATTEMPT++))
+
+  # Show error details
+  echo ""
+  msg_error "VM creation failed in line ${line_number}"
+  msg_error "Exit code: ${exit_code} | Category: ${error_category}"
+  msg_error "Command: ${failed_command}"
+
+  # Show last few lines of error log if available
+  if [[ -s "$VM_ERROR_LOG" ]]; then
+    echo -e "\n${TAB}${YW}--- Last 5 lines of error log ---${CL}"
+    tail -n 5 "$VM_ERROR_LOG" 2>/dev/null | while IFS= read -r line; do
+      echo -e "${TAB}  ${line}"
+    done
+    echo -e "${TAB}${YW}----------------------------------${CL}\n"
+  fi
+
+  # Show recovery menu
+  local choice
+  choice=$(vm_show_recovery_menu "$exit_code" "$error_category" "$VM_RECOVERY_ATTEMPT")
+
+  case "$choice" in
+  RETRY | RETRY_DOWNLOAD)
+    msg_info "Cleaning up failed VM ${VMID} for retry"
+    "$cleanup_fn" 2>/dev/null || true
+    rm -f "$VM_ERROR_LOG"
+    rm -f "${WORK_FILE:-}" 2>/dev/null
+
+    if [[ "$choice" == "RETRY_DOWNLOAD" ]]; then
+      # Clear cached image
+      if [[ -n "${CACHE_FILE:-}" && -f "$CACHE_FILE" ]]; then
+        msg_info "Clearing cached image: $(basename "$CACHE_FILE")"
+        rm -f "$CACHE_FILE"
+        msg_ok "Cache cleared"
+      fi
+    fi
+
+    msg_ok "Ready for retry (attempt $((VM_RECOVERY_ATTEMPT + 1))/${VM_MAX_RETRIES})"
+
+    if [[ -n "$retry_fn" ]]; then
+      # Re-invoke the retry function — caller loop handles this
+      return 0
+    else
+      msg_warn "No retry function provided — please re-run the script manually"
+      return 1
+    fi
+    ;;
+
+  SKIP_CUSTOMIZE)
+    msg_info "Cleaning up failed VM ${VMID} for retry (skipping virt-customize)"
+    "$cleanup_fn" 2>/dev/null || true
+    rm -f "$VM_ERROR_LOG"
+    rm -f "${WORK_FILE:-}" 2>/dev/null
+    # Set flag so docker-vm.sh skips virt-customize
+    export SKIP_VIRT_CUSTOMIZE="yes"
+    msg_ok "Will use first-boot fallback for package installation"
+
+    if [[ -n "$retry_fn" ]]; then
+      return 0
+    else
+      msg_warn "No retry function provided — please re-run the script manually"
+      return 1
+    fi
+    ;;
+
+  RETRY_SETTINGS)
+    msg_info "Cleaning up failed VM ${VMID} for retry with new settings"
+    "$cleanup_fn" 2>/dev/null || true
+    rm -f "$VM_ERROR_LOG"
+    rm -f "${WORK_FILE:-}" 2>/dev/null
+
+    # Let user choose new settings via advanced_settings if available
+    if declare -f advanced_settings >/dev/null 2>&1; then
+      header_info 2>/dev/null || true
+      echo -e "${ADVANCED:-}${BOLD}${RD}Reconfigure VM Settings${CL}"
+      advanced_settings
+    else
+      msg_warn "advanced_settings() not available — using current settings"
+    fi
+
+    if [[ -n "$retry_fn" ]]; then
+      return 0
+    else
+      msg_warn "No retry function provided — please re-run the script manually"
+      return 1
+    fi
+    ;;
+
+  NEW_VMID)
+    msg_info "Cleaning up conflicting VM ${VMID}"
+    "$cleanup_fn" 2>/dev/null || true
+    rm -f "$VM_ERROR_LOG"
+    rm -f "${WORK_FILE:-}" 2>/dev/null
+    VMID=$(get_valid_nextid)
+    echo -e "${CONTAINERID:-}${BOLD}${DGN}New Virtual Machine ID: ${BGN}${VMID}${CL}"
+    msg_ok "Using new VMID: ${VMID}"
+
+    if [[ -n "$retry_fn" ]]; then
+      return 0
+    else
+      msg_warn "No retry function provided — please re-run the script manually"
+      return 1
+    fi
+    ;;
+
+  KEEP)
+    msg_warn "Keeping partial VM ${VMID} for manual debugging"
+    msg_warn "You can inspect it with: qm config ${VMID}"
+    msg_warn "To remove it later: qm destroy ${VMID} --destroy-unreferenced-disks --purge"
+    # Report failure to telemetry
+    post_update_to_api "failed" "$exit_code" 2>/dev/null || true
+    exit "$exit_code"
+    ;;
+
+  ABORT | *)
+    msg_info "Destroying failed VM ${VMID}"
+    "$cleanup_fn" 2>/dev/null || true
+    rm -f "$VM_ERROR_LOG"
+    post_update_to_api "failed" "$exit_code" 2>/dev/null || true
+    msg_error "VM creation aborted by user"
+    exit "$exit_code"
+    ;;
+  esac
+}

vm/debian-13-vm.sh

@@ -65,13 +65,63 @@ trap 'error_handler $LINENO "$BASH_COMMAND"' ERR
 trap cleanup EXIT
 trap 'post_update_to_api "failed" "INTERRUPTED"' SIGINT
 trap 'post_update_to_api "failed" "TERMINATED"' SIGTERM
+
+# Smart recovery state
+VM_CREATION_PHASE="no"
+VM_RECOVERY_ATTEMPT=0
+VM_MAX_RETRIES=2
+
 function error_handler() {
   local exit_code="$?"
   local line_number="$1"
   local command="$2"
   local error_message="${RD}[ERROR]${CL} in line ${RD}$line_number${CL}: exit code ${RD}$exit_code${CL}: while executing command ${YW}$command${CL}"
-  post_update_to_api "failed" "${exit_code}"
   echo -e "\n$error_message\n"
+
+  # During VM creation phase: offer recovery menu instead of immediate cleanup
+  if [[ "$VM_CREATION_PHASE" == "yes" && $VM_RECOVERY_ATTEMPT -lt $VM_MAX_RETRIES ]]; then
+    ((VM_RECOVERY_ATTEMPT++))
+    trap - ERR
+    set +e
+
+    local choice
+    choice=$(whiptail --backtitle "Proxmox VE Helper Scripts" --title "VM CREATION FAILED" \
+      --radiolist "Exit code: ${exit_code} | Attempt: ${VM_RECOVERY_ATTEMPT}/${VM_MAX_RETRIES}\nFailed command: ${command}\n\nChoose a recovery action:" 16 72 4 \
+      "RETRY" "Retry VM creation" "ON" \
+      "SKIP_CUSTOMIZE" "Retry and skip image customization" "OFF" \
+      "KEEP" "Keep partial VM for debugging" "OFF" \
+      "ABORT" "Destroy VM and exit" "OFF" \
+      3>&1 1>&2 2>&3) || choice="ABORT"
+
+    case "$choice" in
+    RETRY | SKIP_CUSTOMIZE)
+      msg_info "Cleaning up failed VM ${VMID} for retry"
+      cleanup_vmid 2>/dev/null || true
+      rm -f "${WORK_FILE:-}" 2>/dev/null
+      [[ "$choice" == "SKIP_CUSTOMIZE" ]] && export SKIP_VIRT_CUSTOMIZE="yes"
+      msg_ok "Ready for retry (attempt $((VM_RECOVERY_ATTEMPT + 1))/${VM_MAX_RETRIES})"
+      set -e
+      trap 'error_handler $LINENO "$BASH_COMMAND"' ERR
+      create_vm
+      exit $?
+      ;;
+    KEEP)
+      echo -e "\n${YW}  Keeping partial VM ${VMID} for debugging${CL}"
+      echo -e "  Inspect: qm config ${VMID}"
+      echo -e "  Remove:  qm destroy ${VMID} --destroy-unreferenced-disks --purge\n"
+      post_update_to_api "failed" "$exit_code"
+      exit "$exit_code"
+      ;;
+    *)
+      post_update_to_api "failed" "$exit_code"
+      cleanup_vmid
+      exit "$exit_code"
+      ;;
+    esac
+  fi
+
+  # Default: no recovery (max retries exceeded or outside creation phase)
+  post_update_to_api "failed" "${exit_code}"
   cleanup_vmid
 }
 
@@ -485,6 +535,7 @@ fi
 msg_ok "Using ${CL}${BL}$STORAGE${CL} ${GN}for Storage Location."
 msg_ok "Virtual Machine ID is ${CL}${BL}$VMID${CL}."
 
+create_vm() {
   # ==============================================================================
   # PREREQUISITES
   # ==============================================================================
@@ -511,11 +562,12 @@ msg_ok "Downloaded ${CL}${BL}${FILE}${CL}"
   # ==============================================================================
   # IMAGE CUSTOMIZATION
   # ==============================================================================
-msg_info "Customizing ${FILE} image"
-
   WORK_FILE=$(mktemp --suffix=.qcow2)
   cp "$FILE" "$WORK_FILE"
 
+  if [[ "${SKIP_VIRT_CUSTOMIZE:-}" != "yes" ]]; then
+    msg_info "Customizing ${FILE} image"
+
     # Set hostname
     virt-customize -q -a "$WORK_FILE" --hostname "${HN}" >/dev/null 2>&1
 
@@ -551,6 +603,9 @@ EOF' >/dev/null 2>&1 || true
     fi
 
     msg_ok "Customized image"
+  else
+    msg_ok "Skipped image customization (hostname and login not pre-configured)"
+  fi
 
   STORAGE_TYPE=$(pvesm status -storage "$STORAGE" | awk 'NR>1 {print $2}')
   case $STORAGE_TYPE in
@@ -650,3 +705,8 @@ fi
 
   msg_ok "Completed successfully!\n"
   echo "More Info at https://github.com/community-scripts/ProxmoxVE/discussions/836"
+} # end create_vm
+
+VM_CREATION_PHASE="yes"
+create_vm
+VM_CREATION_PHASE="no"

vm/docker-vm.sh

@@ -40,10 +40,32 @@ trap cleanup EXIT
 trap 'post_update_to_api "failed" "INTERRUPTED"' SIGINT
 trap 'post_update_to_api "failed" "TERMINATED"' SIGTERM
 
+# Flag to control whether recovery menu is shown (set during create_vm)
+VM_CREATION_PHASE="no"
+
 function error_handler() {
   local exit_code="$?"
   local line_number="$1"
   local command="$2"
+
+  # During VM creation phase: use smart recovery if available
+  if [[ "$VM_CREATION_PHASE" == "yes" ]] && declare -f vm_handle_recovery >/dev/null 2>&1; then
+    # Temporarily disable ERR trap + set -e to prevent recursion during recovery menu
+    trap - ERR
+    set +e
+
+    if vm_handle_recovery "$exit_code" "$line_number" "$command" "cleanup_vmid" "create_vm"; then
+      # Recovery chose retry — re-invoke create_vm with traps restored
+      set -e
+      trap 'error_handler $LINENO "$BASH_COMMAND"' ERR
+      create_vm
+      exit $?
+    fi
+    # Recovery chose abort/keep — vm_handle_recovery already called exit
+    exit "$exit_code"
+  fi
+
+  # Default error handling (outside VM creation phase)
   local error_message="${RD}[ERROR]${CL} in line ${RD}$line_number${CL}: exit code ${RD}$exit_code${CL}: while executing command ${YW}$command${CL}"
   post_update_to_api "failed" "${exit_code}"
   echo -e "\n$error_message\n"
@@ -436,6 +458,15 @@ if ! command -v virt-customize &>/dev/null; then
   msg_ok "Installed libguestfs-tools"
 fi
 
+# ==============================================================================
+# VM CREATION FUNCTION (wrapped for smart recovery retry)
+# ==============================================================================
+create_vm() {
+
+  # Reset error log for this attempt
+  VM_ERROR_LOG="/tmp/vm-install-${VMID}.log"
+  : >"$VM_ERROR_LOG"
+
   # ==============================================================================
   # IMAGE DOWNLOAD
   # ==============================================================================
@@ -492,8 +523,12 @@ export LIBGUESTFS_BACKEND_SETTINGS=dns=8.8.8.8,1.1.1.1
   DOCKER_PREINSTALLED="no"
 
   # Install qemu-guest-agent and Docker during image customization
+  # Skip if recovery set SKIP_VIRT_CUSTOMIZE (virt-customize failed before)
+  if [[ "${SKIP_VIRT_CUSTOMIZE:-}" == "yes" ]]; then
+    msg_ok "Skipping virt-customize (using first-boot fallback)"
+  else
     msg_info "Installing base packages in image"
-if virt-customize -a "$WORK_FILE" --install qemu-guest-agent,curl,ca-certificates >/dev/null 2>&1; then
+    if virt-customize -a "$WORK_FILE" --install qemu-guest-agent,curl,ca-certificates 2>>"$VM_ERROR_LOG" >/dev/null; then
       msg_ok "Installed base packages"
 
       msg_info "Installing Docker (this may take 2-5 minutes)"
@@ -522,6 +557,7 @@ EOF' >/dev/null 2>&1
     else
       msg_ok "Packages will be installed on first boot"
     fi
+  fi
 
   msg_info "Finalizing image (hostname, SSH config)"
   # Set hostname and prepare for unique machine-id
@@ -612,7 +648,7 @@ msg_ok "Resized disk image"
   msg_info "Creating Docker VM shell"
 
   qm create $VMID -agent 1${MACHINE} -tablet 0 -localtime 1 -bios ovmf${CPU_TYPE} -cores $CORE_COUNT -memory $RAM_SIZE \
-  -name $HN -tags community-script -net0 virtio,bridge=$BRG,macaddr=$MAC$VLAN$MTU -onboot 1 -ostype l26 -scsihw virtio-scsi-pci >/dev/null
+    -name $HN -tags community-script -net0 virtio,bridge=$BRG,macaddr=$MAC$VLAN$MTU -onboot 1 -ostype l26 -scsihw virtio-scsi-pci 2>>"$VM_ERROR_LOG" >/dev/null
 
   msg_ok "Created VM shell"
 
@@ -627,7 +663,7 @@ else
     IMPORT_CMD=(qm importdisk)
   fi
 
-IMPORT_OUT="$("${IMPORT_CMD[@]}" "$VMID" "$WORK_FILE" "$STORAGE" ${DISK_IMPORT:-} 2>&1 || true)"
+  IMPORT_OUT="$("${IMPORT_CMD[@]}" "$VMID" "$WORK_FILE" "$STORAGE" ${DISK_IMPORT:-} 2> >(tee -a "$VM_ERROR_LOG") || true)"
   DISK_REF_IMPORTED="$(printf '%s\n' "$IMPORT_OUT" | sed -n "s/.*successfully imported disk '\([^']\+\)'.*/\1/p" | tr -d "\r\"'")"
   [[ -z "$DISK_REF_IMPORTED" ]] && DISK_REF_IMPORTED="$(pvesm list "$STORAGE" | awk -v id="$VMID" '$5 ~ ("vm-"id"-disk-") {print $1":"$5}' | sort | tail -n1)"
   [[ -z "$DISK_REF_IMPORTED" ]] && {
@@ -709,3 +745,13 @@ fi
 
   post_update_to_api "done" "none"
   msg_ok "Completed successfully!\n"
+
+} # end of create_vm()
+
+# ==============================================================================
+# VM CREATION WITH SMART RECOVERY
+# ==============================================================================
+VM_CREATION_PHASE="yes"
+create_vm
+VM_CREATION_PHASE="no"
+rm -f "$VM_ERROR_LOG" 2>/dev/null || true

vm/haos-vm.sh

@@ -69,13 +69,65 @@ trap 'error_handler $LINENO "$BASH_COMMAND"' ERR
 trap cleanup EXIT
 trap 'post_update_to_api "failed" "INTERRUPTED"' SIGINT
 trap 'post_update_to_api "failed" "TERMINATED"' SIGTERM
+
+# Smart recovery state
+VM_CREATION_PHASE="no"
+VM_RECOVERY_ATTEMPT=0
+VM_MAX_RETRIES=2
+
 function error_handler() {
   local exit_code="$?"
   local line_number="$1"
   local command="$2"
   local error_message="${RD}[ERROR]${CL} in line ${RD}$line_number${CL}: exit code ${RD}$exit_code${CL}: while executing command ${YW}$command${CL}"
-  post_update_to_api "failed" "${exit_code}"
   echo -e "\n$error_message\n"
+
+  # During VM creation phase: offer recovery menu instead of immediate cleanup
+  if [[ "$VM_CREATION_PHASE" == "yes" && $VM_RECOVERY_ATTEMPT -lt $VM_MAX_RETRIES ]]; then
+    ((VM_RECOVERY_ATTEMPT++))
+    trap - ERR
+    set +e
+
+    local choice
+    choice=$(whiptail --backtitle "Proxmox VE Helper Scripts" --title "VM CREATION FAILED" \
+      --radiolist "Exit code: ${exit_code} | Attempt: ${VM_RECOVERY_ATTEMPT}/${VM_MAX_RETRIES}\nFailed command: ${command}\n\nChoose a recovery action:" 16 72 4 \
+      "RETRY" "Retry VM creation" "ON" \
+      "RETRY_DOWNLOAD" "Retry with fresh download (clear cache)" "OFF" \
+      "KEEP" "Keep partial VM for debugging" "OFF" \
+      "ABORT" "Destroy VM and exit" "OFF" \
+      3>&1 1>&2 2>&3) || choice="ABORT"
+
+    case "$choice" in
+    RETRY | RETRY_DOWNLOAD)
+      msg_info "Cleaning up failed VM ${VMID} for retry"
+      cleanup_vmid 2>/dev/null || true
+      if [[ "$choice" == "RETRY_DOWNLOAD" && -n "${CACHE_FILE:-}" ]]; then
+        rm -f "$CACHE_FILE"
+        msg_ok "Cleared cached image"
+      fi
+      msg_ok "Ready for retry (attempt $((VM_RECOVERY_ATTEMPT + 1))/${VM_MAX_RETRIES})"
+      set -e
+      trap 'error_handler $LINENO "$BASH_COMMAND"' ERR
+      create_vm
+      exit $?
+      ;;
+    KEEP)
+      echo -e "\n${YW}  Keeping partial VM ${VMID} for debugging${CL}"
+      echo -e "  Inspect: qm config ${VMID}"
+      echo -e "  Remove:  qm destroy ${VMID} --destroy-unreferenced-disks --purge\n"
+      post_update_to_api "failed" "$exit_code"
+      exit "$exit_code"
+      ;;
+    *)
+      post_update_to_api "failed" "$exit_code"
+      cleanup_vmid
+      exit "$exit_code"
+      ;;
+    esac
+  fi
+
+  # Default: no recovery (max retries exceeded or outside creation phase)
+  post_update_to_api "failed" "${exit_code}"
   cleanup_vmid
 }
 
@@ -554,6 +606,7 @@ fi
 msg_ok "Using ${CL}${BL}$STORAGE${CL} ${GN}for Storage Location."
 msg_ok "Virtual Machine ID is ${CL}${BL}$VMID${CL}."
 
+create_vm() {
   var_version="${BRANCH}"
   msg_info "Retrieving the URL for Home Assistant ${BRANCH} Disk Image"
   if [ "$BRANCH" == "$dev" ]; then
@@ -658,3 +711,8 @@ if [ "$START_VM" == "yes" ]; then
   fi
   post_update_to_api "done" "none"
   msg_ok "Completed successfully!\n"
+} # end create_vm
+
+VM_CREATION_PHASE="yes"
+create_vm
+VM_CREATION_PHASE="no"

vm/openwrt-vm.sh

@@ -70,13 +70,61 @@ trap 'error_handler $LINENO "$BASH_COMMAND"' ERR
 trap cleanup EXIT
 trap 'post_update_to_api "failed" "INTERRUPTED"' SIGINT
 trap 'post_update_to_api "failed" "TERMINATED"' SIGTERM
+
+# Smart recovery state
+VM_CREATION_PHASE="no"
+VM_RECOVERY_ATTEMPT=0
+VM_MAX_RETRIES=2
+
 function error_handler() {
   local exit_code="$?"
   local line_number="$1"
   local command="$2"
-  post_update_to_api "failed" "$exit_code"
   local error_message="${RD}[ERROR]${CL} in line ${RD}$line_number${CL}: exit code ${RD}$exit_code${CL}: while executing command ${YW}$command${CL}"
   echo -e "\n$error_message\n"
+
+  # During VM creation phase: offer recovery menu instead of immediate cleanup
+  if [[ "$VM_CREATION_PHASE" == "yes" && $VM_RECOVERY_ATTEMPT -lt $VM_MAX_RETRIES ]]; then
+    ((VM_RECOVERY_ATTEMPT++))
+    trap - ERR
+    set +e
+    set +o pipefail
+
+    local choice
+    choice=$(whiptail --backtitle "Proxmox VE Helper Scripts" --title "VM CREATION FAILED" \
+      --radiolist "Exit code: ${exit_code} | Attempt: ${VM_RECOVERY_ATTEMPT}/${VM_MAX_RETRIES}\nFailed command: ${command}\n\nChoose a recovery action:" 16 72 3 \
+      "RETRY" "Retry VM creation" "ON" \
+      "KEEP" "Keep partial VM for debugging" "OFF" \
+      "ABORT" "Destroy VM and exit" "OFF" \
+      3>&1 1>&2 2>&3) || choice="ABORT"
+
+    case "$choice" in
+    RETRY)
+      msg_info "Cleaning up failed VM ${VMID} for retry"
+      cleanup_vmid 2>/dev/null || true
+      msg_ok "Ready for retry (attempt $((VM_RECOVERY_ATTEMPT + 1))/${VM_MAX_RETRIES})"
+      set -Eeo pipefail
+      trap 'error_handler $LINENO "$BASH_COMMAND"' ERR
+      create_vm
+      exit $?
+      ;;
+    KEEP)
+      echo -e "\n${YW}  Keeping partial VM ${VMID} for debugging${CL}"
+      echo -e "  Inspect: qm config ${VMID}"
+      echo -e "  Remove:  qm destroy ${VMID} --destroy-unreferenced-disks --purge\n"
+      post_update_to_api "failed" "$exit_code"
+      exit "$exit_code"
+      ;;
+    *)
+      post_update_to_api "failed" "$exit_code"
+      cleanup_vmid
+      exit "$exit_code"
+      ;;
+    esac
+  fi
+
+  # Default: no recovery (max retries exceeded or outside creation phase)
+  post_update_to_api "failed" "$exit_code"
   cleanup_vmid
 }
 
@@ -520,6 +568,8 @@ else
 fi
 msg_ok "Using ${CL}${BL}$STORAGE${CL} ${GN}for Storage Location."
 msg_ok "Virtual Machine ID is ${CL}${BL}$VMID${CL}."
+
+create_vm() {
   msg_info "Getting URL for OpenWrt Disk Image"
 
   response=$(curl -fsSL https://openwrt.org)
@@ -661,3 +711,8 @@ if [ -z "$VLAN" ] && [ "$VLAN2" != "999" ]; then
   fi
   post_update_to_api "done" "none"
   msg_ok "Completed Successfully!${VLAN_FINISH:+\n$VLAN_FINISH}"
+} # end create_vm
+
+VM_CREATION_PHASE="yes"
+create_vm
+VM_CREATION_PHASE="no"

vm/ubuntu2204-vm.sh

@@ -62,13 +62,60 @@ trap 'error_handler $LINENO "$BASH_COMMAND"' ERR
 trap cleanup EXIT
 trap 'post_update_to_api "failed" "INTERRUPTED"' SIGINT
 trap 'post_update_to_api "failed" "TERMINATED"' SIGTERM
+
+# Smart recovery state
+VM_CREATION_PHASE="no"
+VM_RECOVERY_ATTEMPT=0
+VM_MAX_RETRIES=2
+
 function error_handler() {
   local exit_code="$?"
   local line_number="$1"
   local command="$2"
-  post_update_to_api "failed" "$exit_code"
   local error_message="${RD}[ERROR]${CL} in line ${RD}$line_number${CL}: exit code ${RD}$exit_code${CL}: while executing command ${YW}$command${CL}"
   echo -e "\n$error_message\n"
+
+  # During VM creation phase: offer recovery menu instead of immediate cleanup
+  if [[ "$VM_CREATION_PHASE" == "yes" && $VM_RECOVERY_ATTEMPT -lt $VM_MAX_RETRIES ]]; then
+    ((VM_RECOVERY_ATTEMPT++))
+    trap - ERR
+    set +e
+
+    local choice
+    choice=$(whiptail --backtitle "Proxmox VE Helper Scripts" --title "VM CREATION FAILED" \
+      --radiolist "Exit code: ${exit_code} | Attempt: ${VM_RECOVERY_ATTEMPT}/${VM_MAX_RETRIES}\nFailed command: ${command}\n\nChoose a recovery action:" 16 72 3 \
+      "RETRY" "Retry VM creation" "ON" \
+      "KEEP" "Keep partial VM for debugging" "OFF" \
+      "ABORT" "Destroy VM and exit" "OFF" \
+      3>&1 1>&2 2>&3) || choice="ABORT"
+
+    case "$choice" in
+    RETRY)
+      msg_info "Cleaning up failed VM ${VMID} for retry"
+      cleanup_vmid 2>/dev/null || true
+      msg_ok "Ready for retry (attempt $((VM_RECOVERY_ATTEMPT + 1))/${VM_MAX_RETRIES})"
+      set -e
+      trap 'error_handler $LINENO "$BASH_COMMAND"' ERR
+      create_vm
+      exit $?
+      ;;
+    KEEP)
+      echo -e "\n${YW}  Keeping partial VM ${VMID} for debugging${CL}"
+      echo -e "  Inspect: qm config ${VMID}"
+      echo -e "  Remove:  qm destroy ${VMID} --destroy-unreferenced-disks --purge\n"
+      post_update_to_api "failed" "$exit_code"
+      exit "$exit_code"
+      ;;
+    *)
+      post_update_to_api "failed" "$exit_code"
+      cleanup_vmid
+      exit "$exit_code"
+      ;;
+    esac
+  fi
+
+  # Default: no recovery (max retries exceeded or outside creation phase)
+  post_update_to_api "failed" "$exit_code"
   cleanup_vmid
 }
 
@@ -466,6 +513,8 @@ else
 fi
 msg_ok "Using ${CL}${BL}$STORAGE${CL} ${GN}for Storage Location."
 msg_ok "Virtual Machine ID is ${CL}${BL}$VMID${CL}."
+
+create_vm() {
   msg_info "Retrieving the URL for the Ubuntu 22.04 Disk Image"
   URL=https://cloud-images.ubuntu.com/jammy/current/jammy-server-cloudimg-amd64.img
   sleep 2
@@ -562,3 +611,8 @@ post_update_to_api "done" "none"
   msg_ok "Completed successfully!\n"
   echo -e "Setup Cloud-Init before starting \n
 More info at https://github.com/community-scripts/ProxmoxVE/discussions/272 \n"
+} # end create_vm
+
+VM_CREATION_PHASE="yes"
+create_vm
+VM_CREATION_PHASE="no"

vm/ubuntu2404-vm.sh

@@ -65,13 +65,60 @@ trap 'error_handler $LINENO "$BASH_COMMAND"' ERR
 trap cleanup EXIT
 trap 'post_update_to_api "failed" "INTERRUPTED"' SIGINT
 trap 'post_update_to_api "failed" "TERMINATED"' SIGTERM
+
+# Smart recovery state
+VM_CREATION_PHASE="no"
+VM_RECOVERY_ATTEMPT=0
+VM_MAX_RETRIES=2
+
 function error_handler() {
   local exit_code="$?"
   local line_number="$1"
   local command="$2"
-  post_update_to_api "failed" "$exit_code"
   local error_message="${RD}[ERROR]${CL} in line ${RD}$line_number${CL}: exit code ${RD}$exit_code${CL}: while executing command ${YW}$command${CL}"
   echo -e "\n$error_message\n"
+
+  # During VM creation phase: offer recovery menu instead of immediate cleanup
+  if [[ "$VM_CREATION_PHASE" == "yes" && $VM_RECOVERY_ATTEMPT -lt $VM_MAX_RETRIES ]]; then
+    ((VM_RECOVERY_ATTEMPT++))
+    trap - ERR
+    set +e
+
+    local choice
+    choice=$(whiptail --backtitle "Proxmox VE Helper Scripts" --title "VM CREATION FAILED" \
+      --radiolist "Exit code: ${exit_code} | Attempt: ${VM_RECOVERY_ATTEMPT}/${VM_MAX_RETRIES}\nFailed command: ${command}\n\nChoose a recovery action:" 16 72 3 \
+      "RETRY" "Retry VM creation" "ON" \
+      "KEEP" "Keep partial VM for debugging" "OFF" \
+      "ABORT" "Destroy VM and exit" "OFF" \
+      3>&1 1>&2 2>&3) || choice="ABORT"
+
+    case "$choice" in
+    RETRY)
+      msg_info "Cleaning up failed VM ${VMID} for retry"
+      cleanup_vmid 2>/dev/null || true
+      msg_ok "Ready for retry (attempt $((VM_RECOVERY_ATTEMPT + 1))/${VM_MAX_RETRIES})"
+      set -e
+      trap 'error_handler $LINENO "$BASH_COMMAND"' ERR
+      create_vm
+      exit $?
+      ;;
+    KEEP)
+      echo -e "\n${YW}  Keeping partial VM ${VMID} for debugging${CL}"
+      echo -e "  Inspect: qm config ${VMID}"
+      echo -e "  Remove:  qm destroy ${VMID} --destroy-unreferenced-disks --purge\n"
+      post_update_to_api "failed" "$exit_code"
+      exit "$exit_code"
+      ;;
+    *)
+      post_update_to_api "failed" "$exit_code"
+      cleanup_vmid
+      exit "$exit_code"
+      ;;
+    esac
+  fi
+
+  # Default: no recovery (max retries exceeded or outside creation phase)
+  post_update_to_api "failed" "$exit_code"
   cleanup_vmid
 }
 
@@ -468,6 +515,8 @@ else
 fi
 msg_ok "Using ${CL}${BL}$STORAGE${CL} ${GN}for Storage Location."
 msg_ok "Virtual Machine ID is ${CL}${BL}$VMID${CL}."
+
+create_vm() {
   msg_info "Retrieving the URL for the Ubuntu 24.04 Disk Image"
   URL=https://cloud-images.ubuntu.com/noble/current/noble-server-cloudimg-amd64.img
   sleep 2
@@ -564,3 +613,8 @@ post_update_to_api "done" "none"
   msg_ok "Completed successfully!\n"
   echo -e "Setup Cloud-Init before starting \n
 More info at https://github.com/community-scripts/ProxmoxVE/discussions/272 \n"
+} # end create_vm
+
+VM_CREATION_PHASE="yes"
+create_vm
+VM_CREATION_PHASE="no"

vm/ubuntu2504-vm.sh

@@ -64,13 +64,60 @@ trap 'error_handler $LINENO "$BASH_COMMAND"' ERR
 trap cleanup EXIT
 trap 'post_update_to_api "failed" "INTERRUPTED"' SIGINT
 trap 'post_update_to_api "failed" "TERMINATED"' SIGTERM
+
+# Smart recovery state
+VM_CREATION_PHASE="no"
+VM_RECOVERY_ATTEMPT=0
+VM_MAX_RETRIES=2
+
 function error_handler() {
   local exit_code="$?"
   local line_number="$1"
   local command="$2"
-  post_update_to_api "failed" "$exit_code"
   local error_message="${RD}[ERROR]${CL} in line ${RD}$line_number${CL}: exit code ${RD}$exit_code${CL}: while executing command ${YW}$command${CL}"
   echo -e "\n$error_message\n"
+
+  # During VM creation phase: offer recovery menu instead of immediate cleanup
+  if [[ "$VM_CREATION_PHASE" == "yes" && $VM_RECOVERY_ATTEMPT -lt $VM_MAX_RETRIES ]]; then
+    ((VM_RECOVERY_ATTEMPT++))
+    trap - ERR
+    set +e
+
+    local choice
+    choice=$(whiptail --backtitle "Proxmox VE Helper Scripts" --title "VM CREATION FAILED" \
+      --radiolist "Exit code: ${exit_code} | Attempt: ${VM_RECOVERY_ATTEMPT}/${VM_MAX_RETRIES}\nFailed command: ${command}\n\nChoose a recovery action:" 16 72 3 \
+      "RETRY" "Retry VM creation" "ON" \
+      "KEEP" "Keep partial VM for debugging" "OFF" \
+      "ABORT" "Destroy VM and exit" "OFF" \
+      3>&1 1>&2 2>&3) || choice="ABORT"
+
+    case "$choice" in
+    RETRY)
+      msg_info "Cleaning up failed VM ${VMID} for retry"
+      cleanup_vmid 2>/dev/null || true
+      msg_ok "Ready for retry (attempt $((VM_RECOVERY_ATTEMPT + 1))/${VM_MAX_RETRIES})"
+      set -e
+      trap 'error_handler $LINENO "$BASH_COMMAND"' ERR
+      create_vm
+      exit $?
+      ;;
+    KEEP)
+      echo -e "\n${YW}  Keeping partial VM ${VMID} for debugging${CL}"
+      echo -e "  Inspect: qm config ${VMID}"
+      echo -e "  Remove:  qm destroy ${VMID} --destroy-unreferenced-disks --purge\n"
+      post_update_to_api "failed" "$exit_code"
+      exit "$exit_code"
+      ;;
+    *)
+      post_update_to_api "failed" "$exit_code"
+      cleanup_vmid
+      exit "$exit_code"
+      ;;
+    esac
+  fi
+
+  # Default: no recovery (max retries exceeded or outside creation phase)
+  post_update_to_api "failed" "$exit_code"
   cleanup_vmid
 }
 
@@ -467,6 +514,8 @@ else
 fi
 msg_ok "Using ${CL}${BL}$STORAGE${CL} ${GN}for Storage Location."
 msg_ok "Virtual Machine ID is ${CL}${BL}$VMID${CL}."
+
+create_vm() {
   msg_info "Retrieving the URL for the Ubuntu 25.04 Disk Image"
   URL=https://cloud-images.ubuntu.com/plucky/current/plucky-server-cloudimg-amd64.img
   sleep 2
@@ -563,3 +612,8 @@ post_update_to_api "done" "none"
   msg_ok "Completed successfully!\n"
   echo -e "Setup Cloud-Init before starting \n
 More info at https://github.com/community-scripts/ProxmoxVE/discussions/272 \n"
+} # end create_vm
+
+VM_CREATION_PHASE="yes"
+create_vm
+VM_CREATION_PHASE="no"