从0开始一步一步搭建CkickHouse集群(单分片双副本)

小咪咪 2023-01-02 13:26 96阅读 0赞

### 文章目录 ###

*   *  CkickHouse 版本
     *  安装包下载
     *  安装
     *  zk集群搭
     *  添加zk配置
     *  测试
     *  常见问题

## CkickHouse 版本 ##

20.3.10.75-2

## 安装包下载 ##

wget https://repo.yandex.ru/clickhouse/rpm/stable/x86_64/clickhouse-client-20.3.10.75-2.noarch.rpm  
    wget https://repo.yandex.ru/clickhouse/rpm/stable/x86_64/clickhouse-server-20.3.10.75-2.noarch.rpm
    wget https://repo.yandex.ru/clickhouse/rpm/stable/x86_64/clickhouse-common-static-dbg-20.3.10.75-2.x86_64.rpm
    wget https://repo.yandex.ru/clickhouse/rpm/stable/x86_64/clickhouse-common-static-20.3.10.75.rpm

也可以使用迅雷加速下载然后上传至服务器  
[下载地址][Link 1]

![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQyNjUxOTA0_size_16_color_FFFFFF_t_70]

![在这里插入图片描述][20210106082139265.png]

## 安装 ##

# server安装
    rpm -ivh clickhouse-common-static-20.3.10.75-2.x86_64.rpm
    rpm -ivh clickhouse-common-static-dbg-20.3.10.75-2.x86_64.rpm
    rpm -ivh clickhouse-server-20.3.10.75-2.noarch.rpm
    
    # client安装
    $ rpm -ivh clickhouse-client-20.3.5.21-2.noarch.rpm

这里的安装仅介绍核心步骤，完整安装步骤可以[参考这里][Link 2]  
集群的搭建和单节点安装没差别，不同的是要配置各个节点与ZooKeeper协同和各个节点的数据同步

## zk集群搭 ##

[参考][Link 3]

## 添加zk配置 ##

在ClickHouse安装完成后，在服务器的/etc/clickhouse-server/config.d目录下创建一个名为metrika.xml的配置文件：

<?xml version="1.0"?>
    <yandex>
        <zookeeper-servers>
          <node index="1">
             <host>ZooKeeper ip</host>
             <port>2181</port>
          </node>
          <node index="2">
             <host>ZooKeeper ip</host>
             <port>2182</port>
          </node>
          <node index="3">
             <host>ZooKeeper ip</host>
             <port>2183</port>
          </node>
       </zookeeper-servers>
    </yandex>

在全局配置config.xml(/etc/clickhouse-server/config.xml)中使用<include\_from>标签导入刚才定义的配置

<include_from>/etc/clickhouse-server/config.d/metrika.xml</include_from>

并引用ZooKeeper配置的定义：

> Zookeeper配置完成我们可以启动服务查看系统提供的zookeeper的代理表  
> SELECT \* FROM system.zookeeper where path = ‘/’  
> ![在这里插入图片描述][20210106083602704.png]

开放外网访问  
取消下面配置的注释

*  配置数据副本同步ip

## 当前clickhouse安装机器ip
    <interserver_http_host>ip</interserver_http_host>

*  集群配置，仅用副本可配可不配做，因为ClickHouse表数据同步仅依赖ZooKeeper就可以了，不依赖集群配置

<remote_servers incl="clickhouse_remote_servers" >
            
            <my_shard>
                   <shard>
    		<internal_replication>true</internal_replication>
                         <replica>
                                <host>node01 ip</host>
                                <port>9000</port>
                              
                         </replica>
                         <replica>
                                <host>node02 ip</host>
                                <port>9001</port>
                         </replica>
                  </shard>
          </my_shard>
          </remote_servers>

> 这里可以看到我只是配置了1个分片一个副本,这里我的一个副本端口使用的是9001，原因是9000端口呗占用了

多副本多分片可以如下配置

完整配置文件参考  
[配置文件][Link 4]

## 测试 ##

第一个节点创建表并插入一条数据

create table t_order
    (
        customer_id Int32 comment '连锁编号',
        order_no String comment '订单号',
        card_no String comment '会员卡号',
        pay_money Decimal64(2) comment '支付金额',
        store_no String comment '门店编号',
        store_user_no String comment '店员编号',
        buy_date Int8 comment '下单时间',
        origin_money Decimal64(2) comment '应收金额',
        year Int16 comment '下单时间-年 2018',
        month Int32 comment '下单时间-月 201808',
        day Int32 comment '下单时间-日 20180801',
        pay_way String comment '支付方式',
        create_date DateTime DEFAULT now() comment '入库时间',
        sign Int8
    )ENGINE = ReplicatedCollapsingMergeTree('/clickhouse/tables/01/t_order','t_order0',sign)
    partition by  (year,month)
    order by (customer_id,year,month,order_no);
    INSERT INTO t_order (customer_id, year, month, order_no, card_no, pay_money, store_no, store_user_no, buy_date, origin_money, day, pay_way, create_date, sign) VALUES (109817, 2020, 202012, '111', '-1', -1.00, '-1', '-1', -1, -1.00, -1, '保留数据', '2020-12-29 10:02:48', -1);

第二个节点创建表

create table t_order
    (
        customer_id Int32 comment '连锁编号',
        order_no String comment '订单号',
        card_no String comment '会员卡号',
        pay_money Decimal64(2) comment '支付金额',
        store_no String comment '门店编号',
        store_user_no String comment '店员编号',
        buy_date Int8 comment '下单时间',
        origin_money Decimal64(2) comment '应收金额',
        year Int16 comment '下单时间-年 2018',
        month Int32 comment '下单时间-月 201808',
        day Int32 comment '下单时间-日 20180801',
        pay_way String comment '支付方式',
        create_date DateTime DEFAULT now() comment '入库时间',
        sign Int8
    )ENGINE = ReplicatedCollapsingMergeTree('/clickhouse/tables/01/t_order','t_order1',sign)
    partition by  (year,month)
    order by (customer_id,year,month,order_no);

要注意 ReplicatedCollapsingMergeTree(’/clickhouse/tables/01/t\_order’,‘t\_order1’,sign) 中 的 zk\_path 是相同的，  
replica\_name 不同  
然后可以看到数据自动同步了，如果从第二个节点插入数据，第一个节点数据也是自动同步的，因为是双主模式，两个节点是对等的

## 常见问题 ##

*  启动错误日志路径:

/var/log/clickhouse-server/clickhouse-server.log

*  查看服务状态

service clickhouse-server status

*  启动

service clickhouse-server start

*  重启

service clickhouse-server restart

*  卸载

yum list installed | grep clickhouse
    yum remove -y clickhouse-common-static
    yum remove -y clickhouse-server-common
    rm -rf /var/lib/clickhouse
    rm -rf /etc/clickhouse-*
    rm -rf /var/log/clickhouse-server

*  [ReplicatedMergeTree 数据同步问题][ReplicatedMergeTree]  
    ![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQyNjUxOTA0_size_16_color_FFFFFF_t_70 1]  
    ![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQyNjUxOTA0_size_16_color_FFFFFF_t_70 2]

[Link 1]: https://repo.yandex.ru/clickhouse/rpm/stable/x86_64/
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQyNjUxOTA0_size_16_color_FFFFFF_t_70]: /images/20221119/4d529b4a085d410192263248fd5a78a8.png
[20210106082139265.png]: /images/20221119/0766f4b5f60549f0b36ecea4ea3b741e.png
[Link 2]: https://weihubeats.blog.csdn.net/article/details/109854977
[Link 3]: https://blog.csdn.net/qq_42651904?spm=1001.2101.3001.5343
[20210106083602704.png]: /images/20221119/c583b1e559d74aaa8f6db3b2f77a6a6b.png
[Link 4]: http://note.youdao.com/noteshare?id=c97c75ba6d97dff445fd1f8e8c8005d0&sub=0882F35755C849ADBEF86682B05324DC
[ReplicatedMergeTree]: http://www.clickhouse.com.cn/topic/5a61957c9d28dfde2ddc5e3a
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQyNjUxOTA0_size_16_color_FFFFFF_t_70 1]: /images/20221119/8308d94adb154ceb9a43a7cd4cd94c7d.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQyNjUxOTA0_size_16_color_FFFFFF_t_70 2]: /images/20221119/de5c1be64adb42d9a5f3c160bc99a06f.png