create_vm

Instructions

Implementation Reference

• src/nutanix_mcp/tools/vm.py:201-235 (handler)

  The async handler function that creates a VM. Extracts arguments (name, cluster_uuid, num_vcpus, memory_mb, disk_size_gb), builds the request body, and calls v4_post to the Nutanix vmm API at 'ahv/config/vms'. Returns a taskExtId for tracking the async VM creation.

  async def handle_create_vm(
      client: NutanixClient, arguments: dict[str, Any]
  ) -> dict[str, Any]:
      """Create a VM using v4 vmm API."""
      name = arguments["name"]
      cluster_uuid = arguments["cluster_uuid"]
      num_vcpus = arguments.get("num_vcpus", 2)
      memory_mb = arguments.get("memory_mb", 4096)
      disk_size_gb = arguments.get("disk_size_gb", 40)
  
      body = {
          "name": name,
          "cluster": {"extId": cluster_uuid},
          "numSockets": 1,
          "numCoresPerSocket": num_vcpus,
          "memorySizeBytes": memory_mb * 1024 * 1024,
          "disks": [
              {
                  "diskSizeBytes": disk_size_gb * 1024 * 1024 * 1024,
                  "storageConfig": {
                      "storageContainerReference": None,
                  },
              }
          ],
      }
  
      result = await client.v4_post(
          namespace="vmm",
          path="ahv/config/vms",
          body=body,
      )
      return {
          "status": "vm_creation_initiated",
          "taskExtId": result.get("data", {}).get("extId"),
      }

• src/nutanix_mcp/tools/vm.py:85-119 (schema)

  The input schema definition for the 'create_vm' tool. Defines parameters: name (required), cluster_uuid (required), num_vcpus (default 2), memory_mb (default 4096), disk_size_gb (default 40).

  {
      "name": "create_vm",
      "description": (
          "Create a new virtual machine. Requires name, cluster UUID, and basic specs."
      ),
      "inputSchema": {
          "type": "object",
          "properties": {
              "name": {
                  "type": "string",
                  "description": "Name for the new VM",
              },
              "cluster_uuid": {
                  "type": "string",
                  "description": "UUID of the cluster to create the VM on",
              },
              "num_vcpus": {
                  "type": "integer",
                  "description": "Number of vCPUs (default: 2)",
                  "default": 2,
              },
              "memory_mb": {
                  "type": "integer",
                  "description": "Memory in MB (default: 4096)",
                  "default": 4096,
              },
              "disk_size_gb": {
                  "type": "integer",
                  "description": "Boot disk size in GB (default: 40)",
                  "default": 40,
              },
          },
          "required": ["name", "cluster_uuid"],
      },
  },

• src/nutanix_mcp/tools/vm.py:240-246 (registration)

  The handler dispatch table mapping the string 'create_vm' to the handle_create_vm function in the VM_HANDLERS dictionary.

  VM_HANDLERS: dict[str, Any] = {
      "list_vms": handle_list_vms,
      "get_vm": handle_get_vm,
      "power_on_vm": handle_power_on_vm,
      "power_off_vm": handle_power_off_vm,
      "create_vm": handle_create_vm,
  }

• src/nutanix_mcp/server.py:34-41 (registration)

  The ALL_HANDLERS dictionary in server.py that merges VM_HANDLERS (among others) into a single dispatch table used by the call_tool handler.

  # Merge all handler dispatch tables
  ALL_HANDLERS: dict[str, Any] = {
      **VM_HANDLERS,
      **CLUSTER_HANDLERS,
      **PE_HANDLERS,
      **REPORT_HANDLERS,
      **NETWORKING_HANDLERS,
  }

• src/nutanix_mcp/server.py:52-62 (registration)

  The list_tools handler in server.py that registers all tool definitions (including create_vm's schema) via the MCP list_tools() decorator.

  @server.list_tools()
  async def list_tools() -> list[Tool]:
      """Return the list of available tools."""
      return [
          Tool(
              name=tool["name"],
              description=tool["description"],
              inputSchema=tool["inputSchema"],
          )
          for tool in all_tools
      ]