watch and build

salt/salt/engines/master/virtual_node_manager.py

@@ -97,8 +97,6 @@ log.setLevel(logging.DEBUG)
 DEFAULT_INTERVAL = 30
 DEFAULT_BASE_PATH = '/opt/so/saltstack/local/salt/hypervisor/hosts'
 VALID_ROLES = ['sensor', 'searchnode', 'idh', 'receiver', 'heavynode', 'fleet']
-MAX_RETRIES = 3
-RETRY_DELAY = 5
 LICENSE_PATH = '/opt/so/saltstack/local/pillar/soc/license.sls'
 
 def ensure_claimed_section(hw_config: dict) -> dict:
@@ -191,17 +189,21 @@ class HardwareManager:
             log.error("Failed to write hypervisor configuration: %s", str(e))
             raise
 
-    def get_pci_ids(self, hw_type: str, indices: List[int]) -> List[str]:
+    def get_pci_ids(self, hw_type: str, indices: Optional[List[int]]) -> List[str]:
         """
         Look up PCI IDs for requested hardware indices.
         
         Args:
             hw_type: Type of hardware (disk, copper, sfp)
-            indices: List of hardware indices to look up
+            indices: List of hardware indices to look up, or None if no hardware requested
         
         Returns:
             List of PCI IDs
         """
+        # Skip if no indices provided
+        if indices is None:
+            return []
+            
         config = self.read_hypervisor_config()
         log.debug("Full config structure: %s", config)
         log.debug("Looking up indices %s for hardware type %s", indices, hw_type)
@@ -343,6 +345,12 @@ def parse_add_file(file_path: str) -> Tuple[str, str, dict]:
             if field not in config:
                 raise ValueError(f"Missing required field: {field}")
         
+        # Validate hardware sections
+        for hw_type in ['disk', 'copper', 'sfp']:
+            if hw_type in config and config[hw_type] is not None:
+                if not isinstance(config[hw_type], list):
+                    raise ValueError(f"{hw_type} must be a list when present")
+                
         return hypervisor, role, config
     except Exception as e:
         log.error("Failed to parse add file %s: %s", file_path, str(e))
@@ -405,6 +413,27 @@ def execute_salt_cloud(profile: str, hostname: str, role: str, config: dict, pci
         log.error("Failed to execute so-salt-cloud: %s", str(e))
         raise
 
+def release_hardware(hw_manager: HardwareManager, hw_type: str, indices: List[int]) -> None:
+    """
+    Release claimed hardware back to free pool.
+    
+    Args:
+        hw_manager: HardwareManager instance
+        hw_type: Type of hardware (disk, copper, sfp)
+        indices: List of hardware indices to release
+    """
+    config = hw_manager.read_hypervisor_config()
+    hw_config = config['hypervisor']['hardware'][hw_type]
+    
+    for idx in indices:
+        if idx in hw_config['claimed']:
+            pci_id = hw_config['claimed'][idx]
+            hw_config['free'][idx] = pci_id
+            del hw_config['claimed'][idx]
+    
+    hw_manager.write_hypervisor_config(config)
+    log.info("Released %s hardware indices: %s", hw_type, indices)
+
 def process_add_file(file_path: str, base_path: str) -> None:
     """
     Process a single add_* file for VM creation.
@@ -418,27 +447,32 @@ def process_add_file(file_path: str, base_path: str) -> None:
         log.debug("Parsed config from add file: %s", config)
         hw_manager = HardwareManager(hypervisor, base_path)
         
-        # Collect all PCI IDs
+        # Phase 1: Collect all hardware information without claiming
         pci_ids = []
+        hardware_to_claim = {}  # Store hardware to claim for each type
         hardware_tracking = {
             'cpu': config.get('cpu'),
             'memory': config.get('memory')
         }
         
-        # Process each hardware type
+        # Validate all hardware first
         for hw_type in ['disk', 'copper', 'sfp']:
-            if hw_type in config:
-                indices = config[hw_type]
+            indices = config.get(hw_type)
+            if indices is not None:
+                # Validate indices are available before claiming
                 hw_pci_ids = hw_manager.get_pci_ids(hw_type, indices)
+                if hw_pci_ids:
+                    hardware_to_claim[hw_type] = indices
                     pci_ids.extend(hw_pci_ids)
-                
-                # Track hardware allocation
                     hardware_tracking[hw_type] = [
                         {'id': idx, 'pci': pci_id}
                         for idx, pci_id in zip(indices, hw_pci_ids)
                     ]
         
-                # Claim the hardware
+        # Phase 2: Claim hardware only after all validation passes
+        try:
+            # Claim all hardware
+            for hw_type, indices in hardware_to_claim.items():
                 hw_manager.claim_hardware(hw_type, indices)
             
             # Create VM
@@ -447,9 +481,27 @@ def process_add_file(file_path: str, base_path: str) -> None:
             # Create hardware tracking file
             hw_manager.create_vm_hardware_file(config['hostname'], role, hardware_tracking)
             
-        # Clean up the add_* file
+            log.info("Successfully processed VM creation request: %s_%s", 
+                    config['hostname'], role)
+            
+        except Exception as e:
+            # If anything fails after claiming, release claimed hardware
+            log.error("Failed after hardware claim, attempting to release hardware: %s", str(e))
+            for hw_type, indices in hardware_to_claim.items():
+                try:
+                    release_hardware(hw_manager, hw_type, indices)
+                except Exception as release_error:
+                    log.error("Failed to release %s hardware %s: %s", 
+                            hw_type, indices, str(release_error))
+            raise
+            
+        finally:
+            # Always clean up the add file
+            try:
                 os.remove(file_path)
-        log.info("Successfully processed VM creation request: %s_%s", config['hostname'], role)
+                log.info("Cleaned up add file: %s", file_path)
+            except Exception as e:
+                log.error("Failed to clean up add file: %s", str(e))
                 
     except Exception as e:
         log.error("Failed to process add file %s: %s", file_path, str(e))
@@ -467,19 +519,10 @@ def monitor_add_files(base_path: str) -> None:
     try:
         for file_path in glob.glob(pattern):
             log.info("Found new VM request file: %s", file_path)
-            
-            for attempt in range(MAX_RETRIES):
             try:
                 process_add_file(file_path, base_path)
-                    break
             except Exception as e:
-                    if attempt < MAX_RETRIES - 1:
-                        log.warning("Attempt %d failed, retrying in %d seconds: %s",
-                                  attempt + 1, RETRY_DELAY, str(e))
-                        time.sleep(RETRY_DELAY)
-                    else:
-                        log.error("All attempts failed for file %s", file_path)
-                        raise
+                log.error("Failed to process file %s: %s", file_path, str(e))
     except Exception as e:
         log.error("Error monitoring add files: %s", str(e))
         raise
@@ -498,9 +541,18 @@ def start(interval: int = DEFAULT_INTERVAL,
     if not validate_hvn_license():
         return
     
+    processed_files = set()
     while True:
         try:
-            monitor_add_files(base_path)
+            for file_path in glob.glob(os.path.join(base_path, '*', 'add_*')):
+                if file_path not in processed_files:
+                    try:
+                        process_add_file(file_path, base_path)
+                        processed_files.add(file_path)
+                    except Exception as e:
+                        log.error("Failed to process file %s: %s", file_path, str(e))
+                        # Still mark as processed to avoid retrying
+                        processed_files.add(file_path)
         except Exception as e:
             log.error("Error in main engine loop: %s", str(e))