diff --git a/.gitignore b/.gitignore
index 9f9225e..7292e9d 100644
--- a/.gitignore
+++ b/.gitignore
@@ -2,4 +2,5 @@
 .DS_Store
 venv
 __pycache__
-env.py
\ No newline at end of file
+env.py
+env.slave.py
\ No newline at end of file
diff --git a/env.py.example b/env.py.example
index c04495c..3b586c2 100644
--- a/env.py.example
+++ b/env.py.example
@@ -70,7 +70,7 @@ QUERY_JOBS = [
         ],
         'allow_less_member': 0,  # 是否允许余票不足时提交部分乘客
         'seats': [  # 筛选座位  有先后顺序 :Array
-            # 可用值: 特等座, 商务座, 一等座, 二等座, 软卧, 硬卧, 硬座, 无座
+            # 可用值: 特等座, 商务座, 一等座, 二等座, 软卧, 硬卧, 动卧, 硬座, 无座
             '硬卧',
             '硬座'
         ],
diff --git a/main.py b/main.py
index 92e6df6..a7918cd 100644
--- a/main.py
+++ b/main.py
@@ -4,15 +4,18 @@ import sys
 from py12306.app import *
 from py12306.log.common_log import CommonLog
 from py12306.query.query import Query
+from py12306.user.user import User
 
 
 def main():
-    if '--test' in sys.argv or '-t' in sys.argv: test()
+    load_argvs()
+    CommonLog.print_welcome()
     App.run()
-    CommonLog.print_welcome().print_configs()
+    CommonLog.print_configs()
     App.did_start()
-    # App.run_check()
-    # User.run()
+
+    App.run_check()
+    User.run()
     Query.run()
     if not Const.IS_TEST:
         while True:
@@ -38,5 +41,15 @@ def test():
     pass
 
 
+def load_argvs():
+    if '--test' in sys.argv or '-t' in sys.argv: test()
+    config_index = None
+
+    if '--config' in sys.argv: config_index = sys.argv.index('--config')
+    if '-c' in sys.argv: config_index = sys.argv.index('-c')
+    if config_index:
+        Config.CONFIG_FILE = sys.argv[config_index + 1:config_index + 2].pop()
+
+
 if __name__ == '__main__':
     main()
diff --git a/py12306/app.py b/py12306/app.py
index bf9eff7..bfa356b 100644
--- a/py12306/app.py
+++ b/py12306/app.py
@@ -40,13 +40,14 @@ class App:
     @classmethod
     def did_start(cls):
         self = cls()
-        if Config.is_cluster_enabled():
-            from py12306.cluster.cluster import Distributed
-            Distributed().join_cluster()
+        # if Config.is_cluster_enabled():
+        #     from py12306.cluster.cluster import Cluster
+        #     Cluster().run()
 
     def init_class(self):
-        from py12306.cluster.cluster import Distributed
-        if Config.is_cluster_enabled(): Distributed()
+        from py12306.cluster.cluster import Cluster
+        if Config.is_cluster_enabled():
+            Cluster().run()
 
     def handler_exit(self, *args, **kwargs):
         """
@@ -56,8 +57,8 @@ class App:
         :return:
         """
         if Config.is_cluster_enabled():
-            from py12306.cluster.cluster import Distributed
-            Distributed().left_cluster()
+            from py12306.cluster.cluster import Cluster
+            Cluster().left_cluster()
 
         sys.exit()
 
diff --git a/py12306/cluster/cluster.py b/py12306/cluster/cluster.py
index c92792f..78f5d9a 100644
--- a/py12306/cluster/cluster.py
+++ b/py12306/cluster/cluster.py
@@ -1,3 +1,8 @@
+import os
+import pickle
+import sys
+import time
+
 import redis
 from redis.client import PubSub
 
@@ -8,60 +13,210 @@ from py12306.log.cluster_log import ClusterLog
 
 
 @singleton
-class Distributed():
+class Cluster():
     KEY_QUERY_COUNT = 'query_count'
     KEY_QUERY_LAST_TIME = 'query_last_time'
     KEY_CONFIGS = 'configs'
     KEY_NODES = 'nodes'
     KEY_CHANNEL_LOG = 'channel_log'
+    KEY_USER_COOKIES = 'user_cookies'
+    KEY_USER_LAST_HEARTBEAT = 'user_last_heartbeat'
+    KEY_NODES_ALIVE = 'nodes_alive'
+
+    KEY_LOCK_INIT_USER = 'lock_init_user'
+    # if self.cluster.get_lock(Cluster.KEY_LOCK_INIT_USER, self.lock_init_user_time):  # TODO 未判断 失败重试
+
+    KEY_MASTER = 1
+    KEY_SLAVE = 0
 
     session: Redis = None
     pubsub: PubSub = None
     refresh_channel_time = 0.5
     retry_time = 2
+    keep_alive_time = 3  # 报告存活间隔
+    lost_alive_time = keep_alive_time * 2
 
+    locks = []
     nodes = {}
+    node_name = None
+    is_ready = False
+    is_master = False
 
     def __init__(self, *args):
-        self.session = Redis()
-        self.pubsub = self.session.pubsub()
-        self.pubsub.subscribe(self.KEY_CHANNEL_LOG)
-        create_thread_and_run(self, 'refresh_data', wait=False)
-        create_thread_and_run(self, 'subscribe', wait=False)
+        if Config.is_cluster_enabled():
+            self.session = Redis()
         return self
 
+    @classmethod
+    def run(cls):
+        self = cls()
+        self.start()
+
+    def start(self):
+        self.pubsub = self.session.pubsub()
+        self.pubsub.subscribe(self.KEY_CHANNEL_LOG)
+        create_thread_and_run(self, 'subscribe', wait=False)
+        self.is_ready = True
+        self.get_nodes()  # 提前获取节点列表
+        self.check_nodes()  # 防止 节点列表未清空
+        self.join_cluster()
+        create_thread_and_run(self, 'keep_alive', wait=False)
+        create_thread_and_run(self, 'refresh_data', wait=False)
+
     def join_cluster(self):
-        node_name = Config().NODE_NAME
+        """
+        加入到集群
+        :return:
+        """
+        self.node_name = node_name = Config().NODE_NAME
+
+        if Config().NODE_IS_MASTER:
+            if self.node_name in self.nodes:  # 重复运行主节点
+                ClusterLog.add_quick_log(ClusterLog.MESSAGE_MASTER_NODE_ALREADY_RUN.format(node_name)).flush(
+                    publish=False)
+                os._exit(1)
+            if self.have_master():  # 子节点提升为主节点情况，交回控制
+                message = ClusterLog.MESSAGE_NODE_BECOME_MASTER_AGAIN.format(node_name)
+                self.publish_log_message(message)
+                self.make_nodes_as_slave()
+        elif not self.have_master():  # 只能通过主节点启动
+            ClusterLog.add_quick_log(ClusterLog.MESSAGE_MASTER_NODE_NOT_FOUND).flush(publish=False)
+            os._exit(1)
+
         if node_name in self.nodes:
-            node_name = node_name + '_' + str(dict_count_key_num(self.nodes, node_name))
+            self.node_name = node_name = node_name + '_' + str(dict_count_key_num(self.nodes, node_name))
             ClusterLog.add_quick_log(ClusterLog.MESSAGE_NODE_ALREADY_IN_CLUSTER.format(node_name)).flush()
 
         self.session.hset(self.KEY_NODES, node_name, Config().NODE_IS_MASTER)
-        message = ClusterLog.MESSAGE_JOIN_CLUSTER_SUCCESS.format(Config().NODE_NAME, list(self.get_nodes()))
-        # ClusterLog.add_quick_log(message).flush()
+        message = ClusterLog.MESSAGE_JOIN_CLUSTER_SUCCESS.format(self.node_name, ClusterLog.get_print_nodes(
+            self.get_nodes()))  # 手动 get nodes
+        self.publish_log_message(message)
+
+    def left_cluster(self, node_name=None):
+        node_name = node_name if node_name else self.node_name
+        self.session.hdel(self.KEY_NODES, node_name)
+        message = ClusterLog.MESSAGE_LEFT_CLUSTER.format(node_name, ClusterLog.get_print_nodes(self.get_nodes()))
+        self.publish_log_message(message)
+
+    def make_nodes_as_slave(self):
+        """
+        将所有节点设为主节点
+        :return:
+        """
+        for node in self.nodes:
+            self.session.hset(self.KEY_NODES, node, self.KEY_SLAVE)
+
+    def publish_log_message(self, message):
+        """
+        发布订阅消息
+        :return:
+        """
+        message = ClusterLog.MESSAGE_SUBSCRIBE_NOTIFICATION.format(self.node_name, message)
         self.session.publish(self.KEY_CHANNEL_LOG, message)
 
-    def left_cluster(self):
-        self.session.hdel(self.KEY_NODES, Config().NODE_NAME)
-        message = ClusterLog.MESSAGE_LEFT_CLUSTER.format(Config().NODE_NAME, list(self.get_nodes()))
-        # ClusterLog.add_quick_log(message).flush()
-        self.session.publish(self.KEY_CHANNEL_LOG, message)
-
-    def get_nodes(self):
+    def get_nodes(self) -> dict:
         res = self.session.hgetall(self.KEY_NODES)
         res = res if res else {}
         self.nodes = res
         return res
 
     def refresh_data(self):
+        """
+        单独进程处理数据同步
+        :return:
+        """
         while True:
             self.get_nodes()
+            self.check_locks()
+            self.check_nodes()
+            self.check_master()
             stay_second(self.retry_time)
 
+    def check_master(self):
+        """
+        检测主节点是否可用
+        :return:
+        """
+        master = self.have_master()
+        if master == self.node_name: # 动态提升
+            self.is_master = True
+        else:
+            self.is_master = False
+
+        if not master:
+            if Config().NODE_SLAVE_CAN_BE_MASTER:
+                # 提升子节点为主节点
+                slave = list(self.nodes)[-1]
+                self.session.hset(self.KEY_NODES, slave, self.KEY_MASTER)
+                self.publish_log_message(ClusterLog.MESSAGE_ASCENDING_MASTER_NODE.format(slave,
+                                                                                         ClusterLog.get_print_nodes(
+                                                                                             self.get_nodes())))
+                return True
+            else:
+                self.publish_log_message(ClusterLog.MESSAGE_MASTER_DID_LOST.format(self.retry_time))
+                stay_second(self.retry_time)
+                os._exit(1)  # 退出整个程序
+
+    def have_master(self):
+        return dict_find_key_by_value(self.nodes, str(self.KEY_MASTER), False)
+
+    def check_nodes(self):
+        """
+        检查节点是否存活
+        :return:
+        """
+        alive = self.session.hgetall(self.KEY_NODES_ALIVE)
+        for node in self.nodes:
+            if node not in alive or (time_int() - int(alive[node])) > self.lost_alive_time:
+                self.left_cluster(node)
+
+    # def kick_out_from_nodes(self, node_name):
+    #     pass
+
+    def keep_alive(self):
+        while True:
+            if self.node_name not in self.get_nodes():  # 已经被 kict out  重新加下
+                self.join_cluster()
+            self.session.hset(self.KEY_NODES_ALIVE, self.node_name, str(time_int()))
+            stay_second(self.keep_alive_time)
+
     def subscribe(self):
         while True:
             message = self.pubsub.get_message()
             if message:
                 if message.get('type') == 'message' and message.get('data'):
-                    ClusterLog.add_quick_log(message.get('data')).flush()
+                    msg = message.get('data')
+                    if self.node_name:
+                        msg = msg.replace(ClusterLog.MESSAGE_SUBSCRIBE_NOTIFICATION_PREFIX.format(self.node_name), '')
+                    ClusterLog.add_quick_log(msg).flush(publish=False)
             stay_second(self.refresh_channel_time)
+
+    def get_lock(self, key, timeout=1):
+        timeout = int(time.time()) + timeout
+        res = self.session.setnx(key, timeout)
+        if res:
+            self.locks.append((key, timeout))
+            return True
+        return False
+
+    def release_lock(self, key):
+        self.session.delete(key)
+
+    def check_locks(self):
+        index = 0
+        for key, timeout in self.locks:
+            if timeout >= int(time.time()):
+                del self.locks[index]
+                self.release_lock(key)
+            index += 1
+
+    @classmethod
+    def get_user_cookie(cls, key, default=None):
+        self = cls()
+        res = self.session.hget(Cluster.KEY_USER_COOKIES, key)
+        return pickle.loads(res.encode()) if res else default
+
+    @classmethod
+    def set_user_cookie(cls, key, value):
+        self = cls()
+        return self.session.hset(Cluster.KEY_USER_COOKIES, key, pickle.dumps(value, 0).decode())
diff --git a/py12306/config.py b/py12306/config.py
index bfe9d3d..3b018b5 100644
--- a/py12306/config.py
+++ b/py12306/config.py
@@ -44,6 +44,7 @@ class Config:
 
     # 集群配置
     CLUSTER_ENABLED = 1
+    NODE_SLAVE_CAN_BE_MASTER = 1
     NODE_IS_MASTER = 1
     NODE_NAME = ''
     REDIS_HOST = ''
@@ -53,6 +54,20 @@ class Config:
     envs = []
     retry_time = 5
 
+    disallow_update_cofigs = [
+        'CLUSTER_ENABLED',
+        'NODE_IS_MASTER',
+        'NODE_NAME',
+        'REDIS_HOST',
+        'REDIS_PORT',
+        'REDIS_PASSWORD',
+    ]
+
+    def __init__(self):
+        self.init_envs()
+        if Config().is_slave():
+            self.refresh_configs(True)
+
     @classmethod
     def run(cls):
         self = cls()
@@ -63,27 +78,26 @@ class Config:
     #     self = cls()
 
     def start(self):
-        self.init_envs()
         self.save_to_remote()
-        # self.refresh_configs()
         create_thread_and_run(self, 'refresh_configs', wait=False)
 
-    def refresh_configs(self):
+    def refresh_configs(self, once=False):
         if not self.is_cluster_enabled(): return
         while True:
             remote_configs = self.get_remote_config()
-            self.update_configs_from_remote(remote_configs)
+            self.update_configs_from_remote(remote_configs, once)
+            if once: break
             stay_second(self.retry_time)
 
     def get_remote_config(self):
         if not self.is_cluster_enabled(): return
-        from py12306.cluster.cluster import Distributed
-        return Distributed().session.get_pickle(Distributed().KEY_CONFIGS, {})
+        from py12306.cluster.cluster import Cluster
+        return Cluster().session.get_pickle(Cluster().KEY_CONFIGS, {})
 
     def save_to_remote(self):
         if not self.is_master(): return
-        from py12306.cluster.cluster import Distributed
-        Distributed().session.set_pickle(Distributed().KEY_CONFIGS, self.envs)
+        from py12306.cluster.cluster import Cluster
+        Cluster().session.set_pickle(Cluster().KEY_CONFIGS, self.envs)
 
     def init_envs(self):
         self.envs = EnvLoader.load_with_file(self.CONFIG_FILE)
@@ -93,31 +107,36 @@ class Config:
         for key, value in envs:
             setattr(self, key, value)
 
-    def update_configs_from_remote(self, envs):
+    def update_configs_from_remote(self, envs, first=False):
         if envs == self.envs: return
         from py12306.query.query import Query
+        from py12306.user.user import User
         for key, value in envs:
-            if key == 'USER_ACCOUNTS' and value != self.USER_ACCOUNTS:  # 用户修改
-                setattr(self, key, value)
-                print('用户修改了') # TODO
-            elif key == 'QUERY_JOBS' and value != self.QUERY_JOBS:  # 任务修改
-                setattr(self, key, value) and Query().update_query_jobs(auto=True)
-            elif key == 'QUERY_INTERVAL' and value != self.QUERY_INTERVAL:  # 任务修改
-                setattr(self, key, value) and Query().update_query_interval(auto=True)
+            if key in self.disallow_update_cofigs: continue
             if value != -1:
+                old = getattr(self, key)
                 setattr(self, key, value)
+                if not first:
+                    if key == 'USER_ACCOUNTS' and old != value:
+                        # 用户修改 print('用户修改了')
+                        User.update_user_accounts(auto=True, old=old)
+                    elif key == 'QUERY_JOBS' and old != value:
+                        Query().update_query_jobs(auto=True)  # 任务修改
+                    elif key == 'QUERY_INTERVAL' and old != value:
+                        Query().update_query_interval(auto=True)
 
     @staticmethod
     def is_master():  # 是不是 主
-        return Config.CLUSTER_ENABLED and Config.NODE_IS_MASTER
+        from py12306.cluster.cluster import Cluster
+        return Config().CLUSTER_ENABLED and (Config().NODE_IS_MASTER or Cluster().is_master)
 
     @staticmethod
     def is_slave():  # 是不是 从
-        return Config.CLUSTER_ENABLED and not Config.NODE_IS_MASTER
+        return Config().CLUSTER_ENABLED and not Config.is_master()
 
     @staticmethod
     def is_cluster_enabled():
-        return Config.CLUSTER_ENABLED
+        return Config().CLUSTER_ENABLED
 
     # @staticmethod
     # def get_members():
diff --git a/py12306/helpers/OCR.py b/py12306/helpers/OCR.py
index 862da08..cef1ff5 100644
--- a/py12306/helpers/OCR.py
+++ b/py12306/helpers/OCR.py
@@ -1,7 +1,7 @@
 import math
 import random
 
-from py12306 import config
+from py12306.config import Config
 from py12306.log.common_log import CommonLog
 from py12306.vender.ruokuai.main import RKClient
 
@@ -22,7 +22,7 @@ class OCR:
         return self.get_img_position_by_ruokuai(img_path)
 
     def get_img_position_by_ruokuai(self, img_path):
-        ruokuai_account = config.AUTO_CODE_ACCOUNT
+        ruokuai_account = Config().AUTO_CODE_ACCOUNT
         soft_id = '119671'
         soft_key = '6839cbaca1f942f58d2760baba5ed987'
         rc = RKClient(ruokuai_account.get('user'), ruokuai_account.get('pwd'), soft_id, soft_key)
@@ -37,7 +37,7 @@ class OCR:
         positions = []
         width = 70
         height = 70
-        random_num = random.randint(0, 10)
+        random_num = random.randint(0, 8)
         for offset in offsets:
             offset = int(offset)
             x = width * (offset % 5) - width / 2 + random_num
diff --git a/py12306/helpers/auth_code.py b/py12306/helpers/auth_code.py
index ef705fb..89753f0 100644
--- a/py12306/helpers/auth_code.py
+++ b/py12306/helpers/auth_code.py
@@ -3,6 +3,7 @@ import time
 
 from requests.exceptions import SSLError
 
+from py12306.config import Config
 from py12306.helpers.OCR import OCR
 from py12306.helpers.api import API_AUTH_CODE_DOWNLOAD, API_AUTH_CODE_CHECK
 from py12306.helpers.request import Request
@@ -20,7 +21,7 @@ class AuthCode:
     retry_time = 5
 
     def __init__(self, session):
-        self.data_path = config.RUNTIME_DIR
+        self.data_path = Config().RUNTIME_DIR
         self.session = session
 
     @classmethod
@@ -66,6 +67,7 @@ class AuthCode:
             UserLog.add_quick_log(UserLog.MESSAGE_CODE_AUTH_SUCCESS).flush()
             return True
         else:
+            # {'result_message': '验证码校验失败', 'result_code': '5'}
             UserLog.add_quick_log(
                 UserLog.MESSAGE_CODE_AUTH_FAIL.format(result.get('result_message'))).flush()
             self.session.cookies.clear_session_cookies()
diff --git a/py12306/helpers/func.py b/py12306/helpers/func.py
index 403db6d..3cd7276 100644
--- a/py12306/helpers/func.py
+++ b/py12306/helpers/func.py
@@ -2,12 +2,12 @@ import datetime
 import random
 import threading
 import functools
+import time
 
 from time import sleep
 from types import MethodType
 
 
-# from py12306 import config
 
 
 def singleton(cls):
@@ -88,11 +88,16 @@ def current_thread_id():
 def time_now():
     return datetime.datetime.now()
 
+def str_to_time(str):
+    return datetime.datetime.strptime(str, '%Y-%m-%d %H:%M:%S.%f')
+
+def time_int():
+    return int(time.time())
+
 
 def create_thread_and_run(jobs, callback_name, wait=True, daemon=True):
     threads = []
-    if not isinstance(jobs, list):
-        jobs = [jobs]
+    if not isinstance(jobs, list): jobs = [jobs]
     for job in jobs:
         thread = threading.Thread(target=getattr(job, callback_name))
         thread.setDaemon(daemon)
@@ -102,6 +107,12 @@ def create_thread_and_run(jobs, callback_name, wait=True, daemon=True):
         for thread in threads: thread.join()
 
 
+def jobs_do(jobs, do):
+    if not isinstance(jobs, list): jobs = [jobs]
+    for job in jobs:
+        getattr(job, do)()
+
+
 def dict_find_key_by_value(data, value, default=None):
     result = [k for k, v in data.items() if v == value]
     return result.pop() if len(result) else default
diff --git a/py12306/helpers/notification.py b/py12306/helpers/notification.py
index f62ed43..c4e7c4d 100644
--- a/py12306/helpers/notification.py
+++ b/py12306/helpers/notification.py
@@ -1,6 +1,5 @@
 import urllib
 
-from py12306 import config
 from py12306.helpers.api import *
 from py12306.helpers.request import Request
 from py12306.log.common_log import CommonLog
diff --git a/py12306/helpers/type.py b/py12306/helpers/type.py
index 2cccc81..c480939 100644
--- a/py12306/helpers/type.py
+++ b/py12306/helpers/type.py
@@ -25,6 +25,7 @@ class OrderSeatType:
         '二等座': 'O',
         '软卧': 4,
         '硬卧': 3,
+        '动卧': 1,
         '硬座': 1,
         '无座': 1,
     }
@@ -39,6 +40,7 @@ class SeatType:
         '二等座': 30,
         '软卧': 23,
         '硬卧': 28,
+        '动卧': 33,
         '硬座': 29,
         '无座': 26,
     }
diff --git a/py12306/log/base.py b/py12306/log/base.py
index d430562..58c4ace 100644
--- a/py12306/log/base.py
+++ b/py12306/log/base.py
@@ -1,5 +1,7 @@
 import os
 import sys
+import io
+from contextlib import redirect_stdout
 
 from py12306.config import Config
 from py12306.helpers.func import *
@@ -23,14 +25,23 @@ class BaseLog:
         return self
 
     @classmethod
-    def flush(cls, sep='\n', end='\n', file=None, exit=False):
+    def flush(cls, sep='\n', end='\n', file=None, exit=False, publish=True):
+        from py12306.cluster.cluster import Cluster
         self = cls()
         logs = self.get_logs()
         # 输出到文件
         if file == None and Config().OUT_PUT_LOG_TO_FILE_ENABLED:  # TODO 文件无法写入友好提示
-            file = open(Config().OUT_PUT_LOG_TO_FILE_PATH, 'a')
+            file = open(Config().OUT_PUT_LOG_TO_FILE_PATH, 'a', encoding='utf-8')
         if not file: file = None
-        print(*logs, sep=sep, end=end, file=file)
+        # 输出日志到各个节点
+        if publish and self.quick_log and Config().is_cluster_enabled() and Cluster().is_ready:  #
+            f = io.StringIO()
+            with redirect_stdout(f):
+                print(*logs, sep=sep, end='' if end == '\n' else end)
+            out = f.getvalue()
+            Cluster().publish_log_message(out)
+        else:
+            print(*logs, sep=sep, end=end, file=file)
         self.empty_logs(logs)
         if exit:
             sys.exit()
diff --git a/py12306/log/cluster_log.py b/py12306/log/cluster_log.py
index 95e0596..0f2000c 100644
--- a/py12306/log/cluster_log.py
+++ b/py12306/log/cluster_log.py
@@ -10,6 +10,26 @@ class ClusterLog(BaseLog):
     quick_log = []
 
     MESSAGE_JOIN_CLUSTER_SUCCESS = '# 节点 {} 成功加入到集群，当前节点列表 {} #'
+
     MESSAGE_LEFT_CLUSTER = '# 节点 {} 已离开集群，当前节点列表 {} #'
 
     MESSAGE_NODE_ALREADY_IN_CLUSTER = '# 当前节点已存在于集群中，自动分配新的节点名称 {} #'
+
+    MESSAGE_SUBSCRIBE_NOTIFICATION_PREFIX = '{} ）'
+    MESSAGE_SUBSCRIBE_NOTIFICATION = MESSAGE_SUBSCRIBE_NOTIFICATION_PREFIX + '{}'
+
+    MESSAGE_ASCENDING_MASTER_NODE = '# 已将 {} 提升为主节点，当前节点列表 {} #'
+
+    MESSAGE_MASTER_DID_LOST = '# 主节点已退出，{} 秒后程序将自动退出 #'
+
+    MESSAGE_MASTER_NODE_ALREADY_RUN = '# 启动失败，主节点 {} 已经在运行中 #'
+    MESSAGE_MASTER_NODE_NOT_FOUND = '# 启动失败，请先启动主节点 #'
+
+    MESSAGE_NODE_BECOME_MASTER_AGAIN = '# 节点 {} 已启动，已自动成功主节点 #'
+
+
+
+    @staticmethod
+    def get_print_nodes(nodes):
+        message = ['{}{}'.format('*' if val == '1' else '', key) for key, val in nodes.items()]
+        return '[ {} ]'.format(', '.join(message))
diff --git a/py12306/log/common_log.py b/py12306/log/common_log.py
index 0819644..acd2ddd 100644
--- a/py12306/log/common_log.py
+++ b/py12306/log/common_log.py
@@ -42,7 +42,7 @@ class CommonLog(BaseLog):
             self.add_quick_log('日志已输出到文件中: {}'.format(Config().OUT_PUT_LOG_TO_FILE_PATH))
 
         self.add_quick_log()
-        self.flush(file=False)
+        self.flush(file=False, publish=False)
         return self
 
     @classmethod
diff --git a/py12306/log/query_log.py b/py12306/log/query_log.py
index ff682a6..6e3fa38 100644
--- a/py12306/log/query_log.py
+++ b/py12306/log/query_log.py
@@ -4,7 +4,7 @@ import sys
 from os import path
 
 from py12306.config import Config
-from py12306.cluster.cluster import Distributed
+from py12306.cluster.cluster import Cluster
 from py12306.log.base import BaseLog
 from py12306.helpers.func import *
 
@@ -36,10 +36,11 @@ class QueryLog(BaseLog):
     def __init__(self):
         super().__init__()
         self.data_path = Config().QUERY_DATA_DIR + 'status.json'
-        self.cluster = Distributed()
-        self.init_data()
+        self.cluster = Cluster()
 
-    def init_data(self):
+    @classmethod
+    def init_data(cls):
+        self = cls()
         # 获取上次记录
         if Const.IS_TEST: return
         result = False
@@ -57,16 +58,16 @@ class QueryLog(BaseLog):
             self.print_data_restored()
 
     def get_data_from_cluster(self):
-        query_count = self.cluster.session.get(Distributed.KEY_QUERY_COUNT, 0)
-        last_time = self.cluster.session.get(Distributed.KEY_QUERY_LAST_TIME, '')
+        query_count = self.cluster.session.get(Cluster.KEY_QUERY_COUNT, 0)
+        last_time = self.cluster.session.get(Cluster.KEY_QUERY_LAST_TIME, '')
         if query_count and last_time:
             return {'query_count': query_count, 'last_time': last_time}
         return False
 
     def refresh_data_of_cluster(self):
         return {
-            'query_count': self.cluster.session.incr(Distributed.KEY_QUERY_COUNT),
-            'last_time': self.cluster.session.set(Distributed.KEY_QUERY_LAST_TIME, time_now()),
+            'query_count': self.cluster.session.incr(Cluster.KEY_QUERY_COUNT),
+            'last_time': self.cluster.session.set(Cluster.KEY_QUERY_LAST_TIME, time_now()),
         }
 
     @classmethod
@@ -144,7 +145,7 @@ class QueryLog(BaseLog):
         self.add_log('=== 正在进行第 {query_count} 次查询 === {time}'.format(query_count=self.data.get('query_count'),
                                                                      time=datetime.datetime.now()))
         if is_main_thread():
-            self.flush()
+            self.flush(publish=False)
         return self
 
     @classmethod
@@ -158,7 +159,7 @@ class QueryLog(BaseLog):
         self.add_quick_log('|=== 查询记录恢复成功 上次查询 {last_date} ===|'.format(last_date=self.data.get('last_time')))
         self.add_quick_log('============================================================')
         self.add_quick_log('')
-        self.flush()
+        self.flush(publish=False)
         return self
 
     def refresh_data(self):
diff --git a/py12306/log/user_log.py b/py12306/log/user_log.py
index e289a15..479754d 100644
--- a/py12306/log/user_log.py
+++ b/py12306/log/user_log.py
@@ -24,6 +24,11 @@ class UserLog(BaseLog):
 
     MESSAGE_WAIT_USER_INIT_COMPLETE = '未找到可用账号或用户正在初始化，{} 秒后重试'
 
+    MESSAGE_USERS_DID_CHANGED = '\n用户信息已更新，正在重新加载...'
+
+    MESSAGE_USER_BEING_DESTROY = '用户 {} 已退出'
+    MESSAGE_USER_COOKIE_NOT_FOUND_FROM_REMOTE = '用户 {} 状态加载中...'
+
     def __init__(self):
         super().__init__()
         self.init_data()
@@ -38,14 +43,14 @@ class UserLog(BaseLog):
         :return:
         """
         self = cls()
-        self.add_log('# 发现 {} 个用户 #'.format(len(users)))
+        self.add_quick_log('# 发现 {} 个用户 #'.format(len(users)))
         self.flush()
         return self
 
     @classmethod
     def print_welcome_user(cls, user):
         self = cls()
-        self.add_log('# 欢迎回来，{} #'.format(user.get_name()))
+        self.add_quick_log('# 欢迎回来，{} #'.format(user.get_name()))
         self.flush()
         return self
 
diff --git a/py12306/query/job.py b/py12306/query/job.py
index 14583e0..2db2460 100644
--- a/py12306/query/job.py
+++ b/py12306/query/job.py
@@ -88,11 +88,11 @@ class Job:
                 self.handle_response(response)
                 self.safe_stay()
                 if is_main_thread():
-                    QueryLog.flush(sep='\t\t')
+                    QueryLog.flush(sep='\t\t', publish=False)
         if is_main_thread():
-            QueryLog.add_quick_log('').flush()
+            QueryLog.add_quick_log('').flush(publish=False)
         else:
-            QueryLog.add_log('\n').flush(sep='\t\t')
+            QueryLog.add_log('\n').flush(sep='\t\t',publish=False)
 
     def query_by_date(self, date):
         """
diff --git a/py12306/query/query.py b/py12306/query/query.py
index 22ff78c..6cfc1cc 100644
--- a/py12306/query/query.py
+++ b/py12306/query/query.py
@@ -1,5 +1,5 @@
 from py12306.config import Config
-from py12306.cluster.cluster import Distributed
+from py12306.cluster.cluster import Cluster
 from py12306.app import app_available_check
 from py12306.helpers.func import *
 from py12306.helpers.request import Request
@@ -23,7 +23,7 @@ class Query:
 
     def __init__(self):
         self.session = Request()
-        self.cluster = Distributed()
+        self.cluster = Cluster()
         self.update_query_interval()
         self.update_query_jobs()
 
@@ -47,6 +47,7 @@ class Query:
     def start(self):
         # return # DEBUG
         self.init_jobs()
+        QueryLog.init_data()
         stay_second(1)
         while True:
             app_available_check()
@@ -64,12 +65,12 @@ class Query:
         QueryLog.print_init_jobs(jobs=self.jobs)
 
     # def get_jobs_from_cluster(self):
-    #     jobs = self.cluster.session.get_dict(Distributed.KEY_JOBS)
+    #     jobs = self.cluster.session.get_dict(Cluster.KEY_JOBS)
     #     return jobs
     #
     # def update_jobs_of_cluster(self):
     #     if config.CLUSTER_ENABLED and config.NODE_IS_MASTER:
-    #         return self.cluster.session.set_dict(Distributed.KEY_JOBS, self.query_jobs)
+    #         return self.cluster.session.set_dict(Cluster.KEY_JOBS, self.query_jobs)
     #
     # def refresh_jobs(self):
     #     if not config.CLUSTER_ENABLED: return
diff --git a/py12306/user/job.py b/py12306/user/job.py
index d596835..853acb1 100644
--- a/py12306/user/job.py
+++ b/py12306/user/job.py
@@ -1,5 +1,7 @@
 import pickle
+from os import path
 
+from py12306.cluster.cluster import Cluster
 from py12306.helpers.api import *
 from py12306.app import *
 from py12306.helpers.auth_code import AuthCode
@@ -10,8 +12,8 @@ from py12306.log.user_log import UserLog
 
 
 class UserJob:
-    heartbeat = 60 * 2  # 心跳保持时长
-    heartbeat_interval = 5
+    # heartbeat = 60 * 2  # 心跳保持时长
+    heartbeat_interval = 60 * 2
     key = None
     user_name = ''
     password = ''
@@ -27,19 +29,29 @@ class UserJob:
     ticket_info_for_passenger_form = None
     order_request_dto = None
 
-    def __init__(self, info, user):
-        self.session = Request()
-        self.heartbeat = user.heartbeat
+    cluster = None
+    lock_init_user_time = 3 * 60
+    cookie = False
 
+    def __init__(self, info):
+        self.cluster = Cluster()
+        self.init_data(info)
+
+    def init_data(self, info):
+        self.session = Request()
         self.key = info.get('key')
         self.user_name = info.get('user_name')
         self.password = info.get('password')
-        self.user = user
+        self.update_user()
+
+    def update_user(self):
+        from py12306.user.user import User
+        self.user = User()
+        self.heartbeat_interval = self.user.heartbeat
+        if not Const.IS_TEST: self.load_user()
 
     def run(self):
         # load user
-        if not Const.IS_TEST:
-            self.load_user()
         self.start()
 
     def start(self):
@@ -49,23 +61,45 @@ class UserJob:
         """
         while True:
             app_available_check()
-            self.check_heartbeat()
+            if Config().is_slave():
+                self.load_user_from_remote()
+            else:
+                if Config().is_master() and not self.cookie: self.load_user_from_remote()  # 主节点加载一次 Cookie
+                self.check_heartbeat()
             if Const.IS_TEST: return
             sleep(self.heartbeat_interval)
 
     def check_heartbeat(self):
         # 心跳检测
-        if self.last_heartbeat and (time_now() - self.last_heartbeat).seconds < self.heartbeat:
+        if self.get_last_heartbeat() and (time_int() - self.get_last_heartbeat()) < self.heartbeat_interval:
             return True
+        # 只有主节点才能走到这
         if self.is_first_time() or not self.check_user_is_login():
             self.handle_login()
 
         self.is_ready = True
-        UserLog.add_quick_log(UserLog.MESSAGE_USER_HEARTBEAT_NORMAL.format(self.get_name(), self.heartbeat)).flush()
+        message = UserLog.MESSAGE_USER_HEARTBEAT_NORMAL.format(self.get_name(), self.heartbeat_interval)
+        if not Config.is_cluster_enabled():
+            UserLog.add_quick_log(message).flush()
+        else:
+            self.cluster.publish_log_message(message)
+        self.set_last_heartbeat()
+
+    def get_last_heartbeat(self):
+        if Config().is_cluster_enabled():
+            return int(self.cluster.session.get(Cluster.KEY_USER_LAST_HEARTBEAT, 0))
+
+        return self.last_heartbeat
+
+    def set_last_heartbeat(self):
+        if Config().is_cluster_enabled():
+            return self.cluster.session.set(Cluster.KEY_USER_LAST_HEARTBEAT, time_int())
         self.last_heartbeat = time_now()
 
     # def init_cookies
     def is_first_time(self):
+        if Config().is_cluster_enabled():
+            return not self.cluster.get_user_cookie(self.key)
         return not path.exists(self.get_cookie_path())
 
     def handle_login(self):
@@ -110,11 +144,10 @@ class UserJob:
 
     def check_user_is_login(self):
         response = self.session.get(API_USER_CHECK.get('url'))
-        is_login = response.json().get('data').get('flag', False)
+        is_login = response.json().get('data.flag', False)
         if is_login:
             self.save_user()
-            self.get_user_info() # 检测应该是不会维持状态，这里再请求下个人中心看有没有有
-
+            self.get_user_info()  # 检测应该是不会维持状态，这里再请求下个人中心看有没有有
 
         return is_login
 
@@ -149,7 +182,7 @@ class UserJob:
         pass
 
     def get_cookie_path(self):
-        return config.USER_DATA_DIR + self.user_name + '.cookie'
+        return Config().USER_DATA_DIR + self.user_name + '.cookie'
 
     def update_user_info(self, info):
         self.info = {**self.info, **info}
@@ -158,6 +191,8 @@ class UserJob:
         return self.info.get('user_name')
 
     def save_user(self):
+        if Config().is_cluster_enabled():
+            return self.cluster.set_user_cookie(self.key, self.session.cookies)
         with open(self.get_cookie_path(), 'wb') as f:
             pickle.dump(self.session.cookies, f)
 
@@ -166,7 +201,7 @@ class UserJob:
         恢复用户成功
         :return:
         """
-        UserLog.add_quick_log(UserLog.MESSAGE_LOADED_USER.format(self.user_name))
+        UserLog.add_quick_log(UserLog.MESSAGE_LOADED_USER.format(self.user_name)).flush()
         if self.check_user_is_login() and self.get_user_info():
             UserLog.add_quick_log(UserLog.MESSAGE_LOADED_USER_SUCCESS.format(self.user_name)).flush()
             UserLog.print_welcome_user(self)
@@ -183,17 +218,42 @@ class UserJob:
         return None
 
     def load_user(self):
+        if Config().is_cluster_enabled(): return
         cookie_path = self.get_cookie_path()
+
         if path.exists(cookie_path):
             with open(self.get_cookie_path(), 'rb') as f:
-                self.session.cookies.update(pickle.load(f))
+                cookie = pickle.load(f)
+                self.cookie = True
+                self.session.cookies.update(cookie)
                 self.did_loaded_user()
                 return True
         return None
 
+    def load_user_from_remote(self):
+        cookie = self.cluster.get_user_cookie(self.key)
+        if not cookie and Config().is_slave():
+            while True:  # 子节点只能取
+                UserLog.add_quick_log(UserLog.MESSAGE_USER_COOKIE_NOT_FOUND_FROM_REMOTE.format(self.user_name)).flush()
+                stay_second(self.retry_time)
+                return self.load_user_from_remote()
+        self.session.cookies.update(cookie)
+        if not self.cookie:  # 第一次加载
+            self.cookie = True
+            self.did_loaded_user()
+        return True
+
     def check_is_ready(self):
         return self.is_ready
 
+    def destroy(self):
+        """
+        退出用户
+        :return:
+        """
+        UserLog.add_quick_log(UserLog.MESSAGE_USER_BEING_DESTROY.format(self.user_name)).flush()
+        sys.exit()
+
     def get_user_passengers(self):
         if self.passengers: return self.passengers
         response = self.session.post(API_USER_PASSENGERS)
diff --git a/py12306/user/user.py b/py12306/user/user.py
index 57b5b1f..89aca10 100644
--- a/py12306/user/user.py
+++ b/py12306/user/user.py
@@ -1,4 +1,5 @@
 from py12306.app import *
+from py12306.cluster.cluster import Cluster
 from py12306.helpers.func import *
 from py12306.log.user_log import UserLog
 from py12306.user.job import UserJob
@@ -8,11 +9,26 @@ from py12306.user.job import UserJob
 class User:
     heartbeat = 60 * 2
     users = []
+    user_accounts = []
 
     retry_time = 3
+    cluster = None
 
     def __init__(self):
-        self.interval = config.USER_HEARTBEAT_INTERVAL
+        self.cluster = Cluster()
+        self.heartbeat = Config().USER_HEARTBEAT_INTERVAL
+        self.update_interval()
+        self.update_user_accounts()
+
+    def update_user_accounts(self, auto=False, old=None):
+        self.user_accounts = Config().USER_ACCOUNTS
+        if auto:
+            UserLog.add_quick_log(UserLog.MESSAGE_USERS_DID_CHANGED).flush()
+            self.refresh_users(old)
+
+    def update_interval(self, auto=False):
+        self.interval = Config().USER_HEARTBEAT_INTERVAL
+        if auto: jobs_do(self.users, 'update_user')
 
     @classmethod
     def run(cls):
@@ -28,17 +44,32 @@ class User:
         create_thread_and_run(jobs=self.users, callback_name='run', wait=Const.IS_TEST)
 
     def init_users(self):
-        accounts = config.USER_ACCOUNTS
-        for account in accounts:
-            user = UserJob(info=account, user=self)
-            self.users.append(user)
+        for account in self.user_accounts:
+            self.init_user(account)
+
+    def init_user(self, info):
+        user = UserJob(info=info)
+        self.users.append(user)
+
+    def refresh_users(self, old):
+        for account in self.user_accounts:
+            key = account.get('key')
+            old_account = array_dict_find_by_key_value(old, 'key', key)
+            if old_account and account != old_account:
+                user = self.get_user(key)
+                user.init_data(account)
+            elif not old_account:
+                self.init_user(account)
+        for account in old:  # 退出已删除的用户
+            if not array_dict_find_by_key_value(self.user_accounts, 'key', account.get('key')):
+                user = self.get_user(account.get('key'))
+                user.destroy()
 
     @classmethod
-    def get_user(cls, key):
+    def get_user(cls, key) -> UserJob:
         self = cls()
         for user in self.users:
-            if user.key == key:
-                return user
+            if user.key == key: return user
         return None
 
     @classmethod