【clickhouse系列】搭建clickhouse集群仅仅需要一个脚本而已

àì夳堔傛蜴生んèń 2022-10-12 00:44 259阅读 0赞

本例子可以支持灵活配置m\*n的clickhouse集群

### 使用方法 ###

分为执行脚本和配置文件两个

#### 配置文件config.toml ####

机器的用户名和密码要配置成一样的

# 分片数
    shard_num = 2
    # 副本数
    replica_num = 2
    # 机器ip列表
    host_list = ["10.0.0.2","10.0.0.3","10.0.0.4","10.0.0.5"]
    # zk ip或域名列表
    zk_host_list = ["10.0.0.6","10.0.0.7","10.0.0.8"]
    # 机器用户名
    user = "root"
    # 机器密码
    password = "xxx"

#### 脚本config\_gen.py ####

import os,sys
    import paramiko
    import toml
    
    # install pip: curl https://bootstrap.pypa.io/get-pip.py | python3
    # install paramiko: pip install paramiko toml
    
    YANDEX = """<yandex> <clickhouse_remote_servers> <default> @cluster@ </default> </clickhouse_remote_servers> <zookeeper-servers> @zookeeper@ </zookeeper-servers> <macros> <shard>@shard@</shard> <replica>@replica@</replica> </macros> </yandex>"""
    
    SHARD = """ <shard> <internal_replication>true</internal_replication> @servers@ </shard> """
    
    REPLICA = """<replica> <host>@host@</host> <port>9000</port> </replica> """
    
    NODE = """<node index="@index@"> <host>@host@</host> <port>2181</port> </node> """
    
    HOST = """<yandex> <listen_host>::</listen_host> <listen_host>0.0.0.0</listen_host> <listen_try>1</listen_try> </yandex> """
    
    # 获取主机名
    def get_hostname(host, user, password):
        ssh = paramiko.SSHClient()
        ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
        ssh.connect(host,22,user,password)
        stdin, stdout, stderr = ssh.exec_command("hostname -s")
        hostname = stdout.read().decode().replace("\n", "")
        return hostname
        
    # 生成cluster相关配置xml节
    def generate_cluster(replica_num, host_list, user, password):
      index = 0
      tmp_shards = ""
      tmp_replica = ""
      while index < len(host_list):
        if index % replica_num == 0:
          if index > 0:
            tmp_shards = tmp_shards + SHARD.replace("@servers@", tmp_replica)
          tmp_replica = ""
          tmp_replica = tmp_replica + REPLICA.replace("@host@", get_hostname(host_list[index], user, password))
        elif index == len(host_list) - 1:
          tmp_replica = tmp_replica + REPLICA.replace("@host@", get_hostname(host_list[index], user, password))
          tmp_shards = tmp_shards + SHARD.replace("@servers@", tmp_replica)
        else:
          tmp_replica = tmp_replica + REPLICA.replace("@host@", get_hostname(host_list[index], user, password))
        index += 1
        
      config = YANDEX.replace("@cluster@", tmp_shards)
      return config
    
    # 生成zk相关配置xml节
    def generate_zookeeper(zk_host_list, config):
      index = 0
      tmp_nodes= ""
      while index < len(zk_host_list):
        node = NODE.replace("@host@", zk_host_list[index])
        node = node.replace("@index@", str(index + 1))
        tmp_nodes = tmp_nodes + node
        index += 1
        
      config = config.replace("@zookeeper@", tmp_nodes)
      return config
    
    # 生成macros相关xml节，并生成各个主机的metrika.xml配置文件，例如metrika.xml-10.0.0.2
    def generate_metrika_config_file(shard_num, replica_num, host_list, config, user, password):
      index = 0
      for i in range(shard_num):
        for j in range(replica_num):
          macros = config.replace("@shard@", str(i))
          macros = macros.replace("@replica@", get_hostname(host_list[index], user, password))
          f = open(sys.path[0] + "/metrika.xml" + "-" + host_list[index], 'w')
          f.write(macros)
          f.close()
          index += 1
    
    # 生成clickhouse允许的ip访问权限配置文件
    def generate_listen_config_file():
      f = open(sys.path[0] + "/listen_host.xml", 'w')
      f.write(HOST)
      f.close()
    
    # 推送metrika.xml-*和listen_host.xml到各个主机
    def push_config_file(host_list, remote_path, user, password):
      if remote_path == "":
        remote_path = "/etc/"
      for host in host_list:
        transport = paramiko.Transport((host, 22))
        transport.connect(username=user, password=password)
        sftp = paramiko.SFTPClient.from_transport(transport)
    
        metrika_path = sys.path[0] + "/metrika.xml-" + host
        sftp.put(metrika_path, remote_path + "metrika.xml")
    
        ssh = paramiko.SSHClient()
        ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
        ssh.connect(host,22,user,password)
        cmd = "mkdir -p /etc/clickhouse-server/config.d"
        stdin, stdout, stderr = ssh.exec_command(cmd)
    
        listen_host_path = sys.path[0] + "/listen_host.xml"
        sftp.put(listen_host_path, remote_path + "clickhouse-server/config.d/listen_host.xml")
        
        transport.close()
    
    # 在各个主机上用yum安装并启动clickhouse
    def install_and_start_clickhouse(host, user, password):
      ssh = paramiko.SSHClient()
      ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
      ssh.connect(host,22,user,password)
      cmd = "{} && {} && {}".format("curl -s https://packagecloud.io/install/repositories/altinity/clickhouse/script.rpm.sh | bash",
                                    "yum install -y clickhouse-server clickhouse-client",
                                    "/etc/init.d/clickhouse-server restart")
      stdin, stdout, stderr = ssh.exec_command(cmd)
      errmsg = stderr.read()
      if errmsg.decode() != "":
        print(errmsg)
      else:
        print("host " + host + " install finished and started")
      ssh.close()
    
    # 获取/etc/hosts需要配置的host信息
    def get_hosts(host_list):
      host_list_str = ""
      for host in host_list:
        ssh = paramiko.SSHClient()
        ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
        ssh.connect(host,22,user,password)
        stdin, stdout, stderr = ssh.exec_command("hostname -s")
        hostname = stdout.read().decode().replace("\n", "")
        stdin, stdout, stderr = ssh.exec_command("ifconfig -a|grep inet|grep -v 127.0.0.1|grep -v inet6|awk '{print $2}'|tr -d \"addr:\"")
        ip = stdout.read().decode().replace("\n", "")
        host_list_str = host_list_str + ip + " " + hostname + "\n"
      return host_list_str
    
    # 将host信息推送到各个主机节点
    def add_hosts(host, host_list_str):
      ssh = paramiko.SSHClient()
      ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
      ssh.connect(host,22,user,password)
      add_host_str = 'echo "{}" >> /etc/hosts'.format(host_list_str)
      stdin, stdout, stderr = ssh.exec_command(add_host_str)
    
    # 启动clickhouse集群
    def start_clickhouse_cluster(host_list, user, password):
      host_list_str = get_hosts(host_list)
      for host in host_list:
        hostname = get_hostname(host, user, password)
        add_hosts(host, "127.0.0.1 " + hostname + "\n" + host_list_str)
        install_and_start_clickhouse(host, user, password)
    
    # 读取配置文件config.toml
    def read_config():
      config_file = 'config.toml'
      if not os.path.exists(config_file):
        input(config_file + ' not found')
        sys.exit(-1)
      with open(config_file, mode='rb') as f:
        content = f.read()
      if content.startswith(b'\xef\xbb\xbf'):
        content = content[3:]
      dic = toml.loads(content.decode('utf8'))
      return dic
    
    
    if __name__ == "__main__":
      dic = read_config()
    
      shard_num = dic['shard_num']
      print('shard_num: %s' % shard_num)
    
      replica_num = dic['replica_num']
      print('replica_num: %s' % replica_num)
    
      user = dic['user'].strip()
      print('user: %s' % user)
    
      password = dic['password'].strip()
      print('password: %s' % password)
    
      host_list = dic['host_list']
      print('host_list: %s' % host_list)
    
      zk_host_list = dic['zk_host_list']
      print('zk_host_list: %s' % zk_host_list)
    
      remote_path = dic['remote_path'].strip()
      print('remote_path: %s' % remote_path)
    
      if len(host_list) % replica_num:
        print("host_list is invalid")
    
      config = generate_cluster(replica_num, host_list, user, password)
      config = generate_zookeeper(zk_host_list, config)
    
      generate_listen_config_file()
    
      generate_metrika_config_file(shard_num, replica_num, host_list, config, user, password)
    
      push_config_file(host_list, remote_path, user, password)
    
      start_clickhouse_cluster(host_list, user, password)

将脚本和其配置文件放在同路径下，运行脚步即可

python3 config_gen.py