Files
FastDeploy/tests/e2e/utils/serving_utils.py
SunLei 782818c031
Some checks failed
CE Compile Job / ce_job_pre_check (push) Has been cancelled
CE Compile Job / print_ce_job_pre_check_outputs (push) Has been cancelled
CE Compile Job / FD-Clone-Linux (push) Has been cancelled
CE Compile Job / Show Code Archive Output (push) Has been cancelled
CE Compile Job / BUILD_SM8090 (push) Has been cancelled
CE Compile Job / BUILD_SM8689 (push) Has been cancelled
CE Compile Job / CE_UPLOAD (push) Has been cancelled
Deploy GitHub Pages / deploy (push) Has been cancelled
fix: ci port conflict (#4840)
2025-11-06 11:56:17 +08:00

97 lines
3.3 KiB
Python

import os
import signal
import socket
import subprocess
import time
# Read ports from environment variables; use default values if not set
FD_API_PORT = int(os.getenv("FD_API_PORT", 8188))
FD_ENGINE_QUEUE_PORT = int(os.getenv("FD_ENGINE_QUEUE_PORT", 8133))
FD_METRICS_PORT = int(os.getenv("FD_METRICS_PORT", 8233))
FD_CACHE_QUEUE_PORT = int(os.getenv("FD_CACHE_QUEUE_PORT", 8333))
# List of ports to clean before and after tests
PORTS_TO_CLEAN = [FD_API_PORT, FD_ENGINE_QUEUE_PORT, FD_METRICS_PORT, FD_CACHE_QUEUE_PORT]
def is_port_open(host: str, port: int, timeout=1.0):
"""
Check if a TCP port is open on the given host.
Returns True if connection succeeds, False otherwise.
"""
try:
with socket.create_connection((host, port), timeout):
return True
except Exception:
return False
def kill_process_on_port(port: int):
"""
Kill processes that are listening on the given port.
Uses multiple methods to ensure thorough cleanup.
"""
current_pid = os.getpid()
parent_pid = os.getppid()
# Method 1: Use lsof to find processes
try:
output = subprocess.check_output(f"lsof -i:{port} -t", shell=True).decode().strip()
for pid in output.splitlines():
pid = int(pid)
if pid in (current_pid, parent_pid):
print(f"Skip killing current process (pid={pid}) on port {port}")
continue
try:
# First try SIGTERM for graceful shutdown
os.kill(pid, signal.SIGTERM)
time.sleep(1)
# Then SIGKILL if still running
os.kill(pid, signal.SIGKILL)
print(f"Killed process on port {port}, pid={pid}")
except ProcessLookupError:
pass # Process already terminated
except subprocess.CalledProcessError:
pass
# Method 2: Use netstat and fuser as backup
try:
# Find processes using netstat and awk
cmd = f"netstat -tulpn 2>/dev/null | grep :{port} | awk '{{print $7}}' | cut -d'/' -f1"
output = subprocess.check_output(cmd, shell=True).decode().strip()
for pid in output.splitlines():
if pid and pid.isdigit():
pid = int(pid)
if pid in (current_pid, parent_pid):
continue
try:
os.kill(pid, signal.SIGKILL)
print(f"Killed process (netstat) on port {port}, pid={pid}")
except ProcessLookupError:
pass
except (subprocess.CalledProcessError, FileNotFoundError):
pass
# Method 3: Use fuser if available
try:
subprocess.run(f"fuser -k {port}/tcp", shell=True, timeout=5)
except (subprocess.TimeoutExpired, subprocess.CalledProcessError, FileNotFoundError):
pass
def clean_ports():
"""
Kill all processes occupying the ports listed in PORTS_TO_CLEAN.
"""
print(f"Cleaning ports: {PORTS_TO_CLEAN}")
for port in PORTS_TO_CLEAN:
kill_process_on_port(port)
# Double check and retry if ports are still in use
time.sleep(2)
for port in PORTS_TO_CLEAN:
if is_port_open("127.0.0.1", port, timeout=0.1):
print(f"Port {port} still in use, retrying cleanup...")
kill_process_on_port(port)
time.sleep(1)