[Feature] [PD Disaggregation] simplify configuration for pd-disaggregated deployment, and refactor post-init and usage for all ports (#5415)

* [feat] simplify configuration for pd-disaggregated deployment, and refactor post-init and usage for all ports

* [fix] fix some bugs

* [fix] fix rdma port for cache manager/messager

* [fix] temporarily cancel port availability check to see if it can pass ci test

* [feat] simplify args for multi api server

* [fix] fix dp

* [fix] fix port for xpu

* [fix] add tests for ports post processing & fix ci

* [test] fix test_multi_api_server

* [fix] fix rdma_comm_ports args for multi_api_server

* [fix] fix test_common_engine

* [fix] fix test_cache_transfer_manager

* [chore] automatically setting FD_ENABLE_MULTI_API_SERVER

* [fix] avoid api server from creating engine_args twice

* [fix] fix test_run_batch

* [fix] fix test_metrics

* [fix] fix splitwise connector init

* [test] add test_rdma_transfer and test_expert_service

* [fix] fix code syntax

* [fix] fix test_rdma_transfer and build wheel with rdma script

fastdeploy/utils.py

@@ -603,6 +603,19 @@ def get_random_port():
                 continue
 
 
+def parse_ports(ports):
+    if ports is None:
+        return None
+    elif isinstance(ports, int):
+        return [ports]
+    elif isinstance(ports, str):
+        return [int(p) for p in ports.split(",")]
+    elif isinstance(ports, list):
+        return [int(p) for p in ports]
+    else:
+        raise TypeError(f"Cannot parse ports into List[int]: {ports}")
+
+
 def is_port_available(host, port):
     """
     Check the port is available
@@ -621,6 +634,57 @@ def is_port_available(host, port):
             return True
 
 
+def find_free_ports(
+    port_range: tuple[int, int] = (8000, 65535),
+    num_ports: int = 1,
+    host: str = "0.0.0.0",
+) -> list[int]:
+    """
+    Find available TCP ports in a given range, scanning from a random start.
+
+    Args:
+        port_range: (start, end), inclusive, e.g. (20000, 30000).
+        num_ports: number of ports to find.
+        host: host to bind, default "0.0.0.0".
+
+    Returns:
+        List of available ports with length == num_ports.
+
+    Raises:
+        ValueError: invalid port range or num_ports <= 0.
+        RuntimeError: not enough free ports in the range.
+    """
+    start, end = port_range
+    if start < 0 or end > 65535 or start > end:
+        raise ValueError(f"Invalid port range: {port_range}")
+
+    if num_ports <= 0:
+        raise ValueError("num_ports must be a positive integer")
+
+    total_ports = end - start + 1
+    if num_ports > total_ports:
+        raise ValueError("num_ports is larger than range size")
+
+    # Generate all ports and rotate with a random start index
+    ports = list(range(start, end + 1))
+    offset = random.randint(0, total_ports - 1)
+    ports = ports[offset:] + ports[:offset]
+
+    free_ports: list[int] = []
+
+    for port in ports:
+        if is_port_available(host, port):
+            free_ports.append(port)
+
+        if len(free_ports) >= num_ports:
+            break
+
+    if len(free_ports) < num_ports:
+        raise RuntimeError(f"Only found {len(free_ports)} free ports in {port_range}, requested {num_ports}.")
+
+    return free_ports
+
+
 def singleton(cls):
     """
     Singleton decorator for a class.