# 实时监控日志 #tail -f /data/server10.log # 查找端口进程 #lsof -i :8000 # 查找server.py进程 #ps aux | grep server.py #source /opt/venv/bin/activate #python a.py huggingface_user="sadww"# huggingface_id="xxrys11" huggingface_tokon="hf_DRZCztytLEC uYjUVsFWIMxqnFgZIiVDERM" ini_content = {"username":"rvbpm6lm","key":"3c17fa4e8261ab5fe7ddea63f2560516"} frp_url = '119.8.118.149' ssh_port = '7082' ssh_url = '45.32.68.209' st_url = f'http://{frp_url}:7092' huggingface_name = 'rvbpm6lm' huggingface_Data_name = 'qilan-st11' data_folder = '/data/SillyTavern' # 要压缩的目录 zip_file_path = '/dev/shm/a/sillytavern.tar.gz' # 输出的 tar.gz 文件路径 metadata_file_path = '/dev/shm/a/dataset-metadata.json' configs = [ (""" [common] server_addr = 47.237.156.222 server_port = 7000 [safsa11] type = tcp local_ip = 127.0.0.1 local_port = 8000 remote_port = 7092 """, '/data/frp/frpc.ini'), (""" [common] server_addr = 47.74.57.131 server_port=7000 [4kxdde11] type=tcp local_ip=127.0.0.1 local_port=8000 remote_port=7092 """, '/data/frp/frpc6.ini'), (""" [common] server_addr = 8.219.236.184 server_port=7000 [safsa11] type=tcp local_ip=127.0.0.1 local_port=8000 remote_port=7092 """, '/data/frp/frpc8.ini'), (""" [common] server_addr = 43.159.47.227 server_port=7000 [safsa11] type=tcp local_ip=127.0.0.1 local_port=8000 remote_port=7092 """, '/data/frp/frpc9.ini'), (""" [common] server_addr = 47.243.129.107 server_port=7000 [4kktyuu11] type=tcp local_ip=127.0.0.1 local_port=8000 remote_port=7092 """, '/data/frp/frpc7.ini') ] kaggle_metadata = { "title": "Your Dataset Title", "id": f"{huggingface_name}/{huggingface_Data_name}", "licenses": [{"name": "CC0-1.0"}] } st_status = True import subprocess import os import sys import configparser # from IPython.display import clear_output import urllib.request from pathlib import Path import requests import json import threading import time import pexpect import socket import time import glob # 导入 glob 模块 from requests.adapters import HTTPAdapter from requests.packages.urllib3.util.retry import Retry import logging from datetime import datetime import psutil from typing import List, Dict import pytz # 创建 FRP 文件夹并下载 frpc 可执行文件。 os.system("wget -q https://huggingface.co/datasets/Qilan2/st-server/raw/main/ip_huggingface.py -O /data/ip_huggingface.py") os.system("wget -q https://huggingface.co/datasets/Qilan2/st-server/resolve/main/ipn.py?download=true -O /data/ipn.py") os.system("wget -q https://huggingface.co/datasets/Qilan2/st-server/resolve/main/hf-jg/11.py?download=true -O /data/sbx.py") os.system("mkdir -p /data/frp") os.system("wget /data/frp/frpc -q https://huggingface.co/Qilan2/box/resolve/main/frp/frpc -O /data/frp/frpc") os.system("chmod +x /data/frp/frpc") os.system("wget -q https://huggingface.co/datasets/Qilan2/st-server/resolve/main/xugou-agent-linux-amd64?download=true -O /data/xugou-agent") os.system("chmod +x /data/xugou-agent") # os.system('nohup /data/xugou-agent start --server "https://tz.282820.xyz" --token "xugou_meqz1qus_80244ac171f1d9cf3a1ead6be2ecf86f.7d4ae786cbb7e559" --interval "30" > /data/xugou-agent.log 2>&1 &') # os.system("wget -O '/data/nezha-agent-amd64' 'https://huggingface.co/datasets/Qilan2/st-server/resolve/main/nezha-agent-v0.17.6?download=true'") # os.system("chmod +x /data/nezha-agent-amd64") # os.system('nohup /data/nezha-agent-amd64 -s tzz.282820.xyz:443 -p aLLlygYqAXgjq9Anln --tls --debug --temperature --disable-force-update --disable-auto-update > nezha-agent-amd64.log 2>&1 &') # os.system('/data/nezha-agent-amd64 -s tzz.282820.xyz:443 -p aLLlygYqAXgjq9Anln --tls --debug --temperature --disable-force-update --disable-auto-update') import yaml def nezha(): agent_config ={ 'client_secret': 'MLcD6YnifhoY08B9n129UP5cg2139NYa', 'debug': False, 'disable_auto_update': True, 'disable_command_execute': False, 'disable_force_update': False, 'disable_nat': False, 'disable_send_query': False, 'gpu': False, 'insecure_tls': False, 'ip_report_period': 1800, 'report_delay': 3, 'self_update_period': 0, 'server': 'z.282820.xyz:443', 'skip_connection_count': False, 'skip_procs_count': False, 'temperature': True, 'tls': True, 'use_gitee_to_upgrade': False, 'use_ipv6_country_code': False, 'uuid': '172528d0-a303-4096-b8d3-df117b8e1c4b' } with open('./config.yaml', 'w') as file: yaml.dump(agent_config, file, default_flow_style=False) os.system("rm -rf f1") os.system("wget -O 'f1' -q 'https://huggingface.co/datasets/Qilan2/st-server/resolve/main/nv1amd64?download=true'") os.system("chmod +x f1") os.system('./f1 -c ./config.yaml') os.system('./f1 -c ./config.yaml') def cloudflared(): os.system("rm -r /data/cloudflared-linux-amd64") os.system("wget -O '/data/cloudflared-linux-amd64' -q 'https://github.com/cloudflare/cloudflared/releases/download/2025.9.0/cloudflared-linux-amd64'") os.system("chmod +x /data/cloudflared-linux-amd64") os.system('/data/cloudflared-linux-amd64 tunnel run --token eyJhIjoiNmVjNWE5NzEzOGIzMTg4YTU2Y2U1NjdmMWRhZDBhMTUiLCJ0IjoiMzg0OWFhNjgtMDJhMS00YWYyLWEzMTMtYzQ3ZmRiYTRjNzE3IiwicyI6IlpUTTRZemRsT1RFdE5tVXhOUzAwTlRVeUxUbGhZbUV0TmpFM09UWm1NR1ExTkdJMCJ9') def sbx(): os.system("python /data/sbx.py") threading.Thread(target=nezha, daemon=True).start() def kill_frpc_processes(): try: # 使用pkill命令结束所有frpc进程 subprocess.run(['pkill', '-f', 'frpc'], check=True) print("所有frpc进程已被结束。") except subprocess.CalledProcessError as e: # 如果pkill命令失败，打印错误信息 print(f"结束frpc进程时出错: {e}") kill_frpc_processes() # 创建一个锁 lock = threading.Lock() last_uploaded_size = -1 def update_last_uploaded_size(size): global last_uploaded_size last_uploaded_size = size def get_folder_size(folder): """使用 du 命令获取文件夹的总大小（字节数）""" result = subprocess.run(['du', '-sb', folder], stdout=subprocess.PIPE, stderr=subprocess.PIPE) if result.returncode == 0: size = result.stdout.split()[0] # 获取大小 return int(size) else: print(f"Error calculating size: {result.stderr.decode()}") return 0 def compress_folder(): global last_uploaded_size # 声明使用全局变量 # 检查是否是中国时区早上5点 china_tz = pytz.timezone('Asia/Shanghai') now = datetime.now(china_tz) if now.hour == 5: print(f"当前时间: {now.strftime('%Y-%m-%d %H:%M:%S')}, 早上5点，跳过备份操作") return # 清理并创建目录 /a if os.path.exists('/dev/shm/a'): subprocess.run(['rm', '-rf', '/dev/shm/a']) os.makedirs('/dev/shm/a', exist_ok=True) with lock: # 使用锁 # 获取当前文件夹的大小 current_folder_size = get_folder_size(data_folder) print(f"当前文件夹大小: {current_folder_size} 字节") # 判断文件夹大小是否大于 800MB (800MB = 800 * 1024 * 1024 字节) if current_folder_size < (200 * 1024 ** 2): print("文件夹小于800MB，不进行上传和压缩。") return # 如果是首次运行或者文件夹大小不同于上次上传，进行压缩 if last_uploaded_size == -1 or current_folder_size != last_uploaded_size: print('开始压缩数据') try: # 执行压缩 result = subprocess.run( # tar -cvf - /data/SillyTavern | pigz -p 2 -1 > /dev/shm/a/sillytavern.tar.gz 2> /dev/null f'tar -cvf - {data_folder} | pigz -p 2 -1 > {zip_file_path} 2> /dev/null', shell=True, check=True, stdout=subprocess.DEVNULL, # 将标准输出重定向到 /dev/null stderr=subprocess.DEVNULL # 将标准错误也重定向到 /dev/null ) except subprocess.CalledProcessError as e: print(f"命令失败，返回代码: {e.returncode}") return print('压缩完成') file_path = "/dev/shm/a/sillytavern.ctk" subprocess.run(['mv', zip_file_path, file_path]) print(f"数据压缩为 {file_path}") file_size = os.path.getsize(file_path) # 获取压缩包的大小 print('文件大小:', file_size) if file_size < (200 * 1024 ** 2): # 判断压缩后的大小是否小于 800MB print("压缩文件小于800MB，不进行上传。") return # 执行上传 if last_uploaded_size == -1: print("首次运行，正在执行上传...") update_last_uploaded_size(file_size) with open(metadata_file_path, 'w') as json_fid: json.dump(kaggle_metadata, json_fid) subprocess.run(['kaggle', 'datasets', 'version', '-m', 'new version message', '-p', '/dev/shm/a']) #kaggle datasets version -m 'new version message' -p /dev/shm/a last_uploaded_size = file_size else: print(f"新文件大小 {file_size} 大于上次上传的大小，正在执行上传...") update_last_uploaded_size(file_size) with open(metadata_file_path, 'w') as json_fid: json.dump(kaggle_metadata, json_fid) subprocess.run(['kaggle', 'datasets', 'version', '-m', 'new version message', '-p', '/dev/shm/a']) last_uploaded_size = file_size else: print(f"新文件夹大小 {current_folder_size} 不大于上次上传的大小，跳过") def repeat_task(): print('------打包线程启动-------------') while True: print('等待打包') # time.sleep(600) time.sleep(10800)# 2小时 compress_folder() def requests_retry_session(retries=3, backoff_factor=0.3, status_forcelist=(500, 502, 504), session=None): """创建一个带重试机制的请求会话""" session = session or requests.Session() retry = Retry(total=retries, backoff_factor=backoff_factor, status_forcelist=status_forcelist) adapter = HTTPAdapter(max_retries=retry) session.mount('http://', adapter) session.mount('https://', adapter) return session def kill_processes_on_port(port: int = 8000) -> Dict[str, List[int]]: """ 结束指定端口的所有相关进程 Args: port (int): 要处理的端口号，默认为8000 Returns: Dict[str, List[int]]: 结束的进程信息 """ killed_processes = { 'success': [], 'failed': [] } try: # 遍历所有进程 for proc in psutil.process_iter(['pid', 'name', 'cmdline']): try: # 检查进程的连接 connections = proc.net_connections() # 检查是否有指定端口的连接 port_matches = [ conn for conn in connections if conn.laddr.port == port ] if port_matches: pid = proc.pid name = proc.name() print(f"找到端口 {port} 相关进程: PID {pid}, 名称 {name}") try: # 尝试优雅地终止进程 parent = psutil.Process(pid) # 终止进程及其所有子进程 for child_proc in parent.children(recursive=True): try: child_proc.terminate() print(f"终止子进程: PID {child_proc.pid}") except Exception as child_err: print(f"无法终止子进程 {child_proc.pid}: {child_err}") parent.terminate() # 等待进程结束 gone, alive = psutil.wait_procs([parent], timeout=3) for p in alive: print(f"强制杀死进程: PID {p.pid}") p.kill() killed_processes['success'].append(pid) print(f"成功终止进程 {pid}") except psutil.NoSuchProcess: print(f"进程 {pid} 已不存在") except psutil.AccessDenied: print(f"没有权限终止进程 {pid}") killed_processes['failed'].append(pid) except Exception as e: print(f"终止进程 {pid} 时发生错误: {e}") killed_processes['failed'].append(pid) except (psutil.NoSuchProcess, psutil.AccessDenied): continue except Exception as e: print(f"搜索进程时发生错误: {e}") # 打印总结报告 print("--- 进程终止报告 ---") print(f"成功终止进程数: {len(killed_processes['success'])}") print(f"终止失败进程数: {len(killed_processes['failed'])}") if killed_processes['success']: print(f"成功终止的进程 PID: {killed_processes['success']}") if killed_processes['failed']: print(f"终止失败的进程 PID: {killed_processes['failed']}") return killed_processes def check_api1(): # 第一次检查 SillyTavern 是否启动 thread_started = False n = 0 while True: n += 1 if not os.path.exists(data_folder): logging.error(f"/sillytavern/ 文件不存在，结束检测") break try: url = "http://127.0.0.1:8000/v1" response = requests.get(url, timeout=30) if response.status_code == 403: print("SillyTavern已经启动") break print("SillyTavern未启动，状态码:", response.status_code, n) except requests.exceptions.RequestException as e: print(f"SillyTavern未启动3: {e}, 尝试次数: {n}") time.sleep(5) print('ST本地已经运行') def frp(configs, backstage): print('-----------FRP开始-------------') # 遍历所有配置并启动 FRP。 for config_data, config_file_path in configs: with open(config_file_path, 'w') as config_file: config_file.write(config_data) # 启动 frpc 命令和参数。 frpc_cmd=['/data/frp/frpc', '-c', config_file_path] if backstage: subprocess.Popen(frpc_cmd) else: subprocess.run(f'/data/frp/frpc -c {config_file_path}', shell=True, check=True) print('-----------FRP结束-------------') def monitor_port(): while True: url = "http://127.0.0.1:8000/v1" response = requests.get(url, timeout=30) if response.status_code != 403: print("SillyTavern未运行，状态码:", response.status_code, n) # 读取 config.ini 中的 monitor_port_enabled 配置项 config = configparser.ConfigParser() config.read('/data/config.ini') monitor_port_enabled = config.getboolean('general', 'monitor_port_enabled', fallback=False) if monitor_port_enabled: kill_processes_on_port() backup_thread = threading.Thread(target=Run_SillyTavern, daemon=True) backup_thread.start() else: print(f"monitor_port_enabled 为False跳过") time.sleep(15) # 每5秒检查一次 def Run_SillyTavern(): try: os.chdir('/data/SillyTavern') # 启动 Node.js 服务器，并将输出重定向到 /dev/null print('-----启动st-----') os.system("yes | sh ./start.sh") os.system("yes | sh ./start.sh") os.system("yes | sh ./start.sh") os.system("yes | sh ./start.sh") os.system("yes | sh ./start.sh") # subprocess.run(f'yes | sh ./start.sh', shell=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL, check=True) # subprocess.run(f'yes | sh {data_folder}/start.sh', shell=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL, check=True) # subprocess.run(f'yes | sh {data_folder}/start.sh', shell=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL, check=True) # subprocess.run(f'yes | sh {data_folder}/start.sh', shell=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL, check=True) # subprocess.run(f'yes | sh {data_folder}/start.sh', shell=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL, check=True) print("已尝试启动 server.js") except subprocess.CalledProcessError as e: print(f"命令失败，返回码: {e.returncode}") except Exception as e: # 捕获所有其他异常 print(f"发生异常: {e}") def SillyTavern(): print('-----------SillyTavern开始-------------') # 创建目录并写入 kaggle.json # 打印当前工作目录和用户主目录 print("Current Working Directory:", os.getcwd()) print("Home Directory:", os.path.expanduser('~')) kaggle_dir = os.path.expanduser('~/.kaggle') kaggle_file_path = os.path.join(kaggle_dir, 'kaggle.json') # 确保目录存在 os.makedirs(kaggle_dir, exist_ok=True) # 显式写入并验证 try: with open(kaggle_file_path, 'w') as json_file: json.dump(ini_content, json_file, indent=4) # 验证文件内容 with open(kaggle_file_path, 'r') as json_file: print("文件内容:", json_file.read()) # 设置权限 os.chmod(kaggle_file_path, 0o600) print(f"文件已写入: {kaggle_file_path}") except Exception as e: print(f"发生错误: {e}") os.system(f"rm -rf /a /data/sillytavern /data/data /data/*.zip /data/*.ctk") # os.system(f"mkdir -p {data_folder}") print('----当前文件夹路径--------') os.system('pwd') print('----进入/data--------') os.chdir('/data') print(f"kaggle datasets download {huggingface_name}/{huggingface_Data_name}") # os.system(f"kaggle datasets download {huggingface_name}/{huggingface_Data_name}") download_n = 0 while True: # 执行下载命令 if download_n > 6: break os.system(f"kaggle datasets download {huggingface_name}/{huggingface_Data_name} -p /data") # 检查文件是否存在 if os.path.exists(f'/data/{huggingface_Data_name}.zip'): print(f"{huggingface_Data_name}.zip 下载成功！") break # 跳出循环 print(f"{huggingface_Data_name}.zip 下载失败，正在重试...") download_n += 1 time.sleep(300) # 等待一段时间再重试（例如5秒 # os.system('kaggle datasets download xovale7506/sillytavern-data2') # 检查文件是否存在 if not os.path.exists(f'/data/{huggingface_Data_name}.zip'): print(f"文件 /{huggingface_Data_name}.zip 不存在，启用实时同步。") else: os.chdir('/data') # 如果文件存在，可以继续执行后续代码 print('----/data文件夹内容--------') # os.system("ls") print(f"文件 /data/{huggingface_Data_name}.zip 存在，继续执行程序。") # # 解压 sillytavern.zip # os.system("rm *.zip") os.system(f"7z x /data/{huggingface_Data_name}.zip -y") # os.system("mv /data/sillytavern.tck sillytavern.tar.gz") os.system("mv /data/sillytavern.ctk sillytavern.tar.gz") # os.system("ls /data") # os.system("tar -xvpf ./sillytavern.tar.gz -C / --use-compress-program=pigz") # 使用标准 tar 解压 os.system("tar -xzf sillytavern.tar.gz --strip-components=1 -C /data") # os.system("pv /sillytavern.tar.gz | tar -xvpf - -C / --use-compress-program=pigz") # os.system("mv /data/data/SillyTavern /data/SillyTavern") print('更新npm') print('----进入/data/SillyTavern--------') os.chdir('/data/SillyTavern') os.system("pwd") # print('----/data/SillyTavern内容--------') os.system(f"ls {data_folder}") os.system("nvm install node") os.system("nvm use node") os.system("npm install -y") # os.system("yes | sh ./start.sh") # try: # subprocess.run(f'yes | sh {data_folder}/start.sh', shell=True, check=True) # except subprocess.CalledProcessError as e: # print(f"Command '{e.cmd}' returned non-zero exit status {e.returncode}.") # print("Output:", e.output) # os.system(f"ls {data_folder}") # Run_SillyTavern() backup_thread = threading.Thread(target=Run_SillyTavern, daemon=True) backup_thread.start() # os.system(f"rm -rf /a /data/sillytavern /data/data /data/*.zip /data/*.ctk") # os.system("ls /data/") print('-----------SillyTavern结束-------------') import requests def _reconstruct_token(partial_token): """ 重构完整的 token :param partial_token: 部分 token :return: 完整的 token """ return partial_token.replace(" ", "") def restart_huggingface_space(space_name, space_id, partial_token): """ 重启 Hugging Face Space :param space_name: Space 的命名空间 :param space_id: Space 的 ID :param partial_token: Hugging Face 部分访问令牌 :return: 响应结果字典 """ # 重构完整 token token = _reconstruct_token(partial_token) url = f"https://huggingface.co/api/spaces/{space_name}/{space_id}/restart?factory=true" headers = { "Content-Type": "application/json", "Authorization": f"Bearer {token}", "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/137.0.0.0 Safari/537.36" } try: response = requests.post(url, headers=headers, json={}) return { "status_code": response.status_code, "success": response.status_code == 200, "message": response.text } except requests.RequestException as e: return { "status_code": None, "success": False, "message": str(e) } # 使用示例 import psutil def check_system_resources(): print("查看CPU占用情况") time.sleep(120) # 获取CPU使用率 cpu_usage = psutil.cpu_percent(interval=1) # 获取内存使用率 memory = psutil.virtual_memory() memory_usage = memory.percent # 检查CPU和内存占用是否超过85% # if cpu_usage >= 90: if cpu_usage >= 90 or memory_usage >= 95: print("占用过高") # 可选：打印具体的使用率 print(f"CPU使用率: {cpu_usage}%") print(f"内存使用率: {memory_usage}%") result = restart_huggingface_space(huggingface_user, huggingface_id, huggingface_tokon) print(result) # time.sleep(6666666) else: print("系统资源正常") # 调用函数 threading.Thread(target=SillyTavern, daemon=True).start() # SillyTavern() time.sleep(60) check_api1() # compress_folder() # 创建 ConfigParser 对象 config = configparser.ConfigParser() # 添加一个 'general' 部分 config.add_section('general') # 设置 'monitor_port_enabled' 选项的值为 True config.set('general', 'monitor_port_enabled', 'True') # 将配置写入到 'config.ini' 文件中 with open('/data/config.ini', 'w') as configfile: config.write(configfile) if os.path.exists(f'/data/{huggingface_Data_name}.zip'): os.system(f"rm -rf /a /data/data /data/*.zip /data/*.ctk /data/*.gz") backup_thread = threading.Thread(target=repeat_task, daemon=True)#压缩备份 backup_thread.start() print("开始衔接") Service_status = True frp(configs, True) threading.Thread(target=monitor_port, daemon=True).start()#检测端口状态 threading.Thread(target=check_system_resources, daemon=True).start()#检测CPU内存状态 threading.Thread(target=sbx, daemon=True).start() os.system("python /data/ip_huggingface.py") os.system("python /data/ipn.py") # time.sleep(600) while True: # 检查是否是中国时区早上6点 china_tz = pytz.timezone('Asia/Shanghai') now = datetime.now(china_tz) if now.hour == 6 and now.minute < 5: # 6:00-6:05之间退出 print(f"当前时间: {now.strftime('%Y-%m-%d %H:%M:%S')}, 到达早上6点，退出循环") break # else: # print(f"当前时间: {now.strftime('%Y-%m-%d %H:%M:%S')}, 未到达早上6点") if backup_thread.is_alive() == False: print("备份线程已停止.") backup_thread = threading.Thread(target=repeat_task, daemon=True) backup_thread.start() time.sleep(5) else: print(f'{data_folder}/start.sh 文件不存在结束执行') compress_folder() result = restart_huggingface_space(huggingface_user, huggingface_id, huggingface_tokon) print(result)