ClickHouse安装使用(单机、集群、高可用)

时间:2021-01-18 小跑、追幸福人气:0

Clickhouse版本：20.3.6.40-2

安装包地址：https://repo.yandex.ru/clickhouse/rpm/stable/x86_64/

一、单机版

1、安装依赖

yum install libicu.x86_64

2、下载安装包

http://repo.yandex.ru/clickhouse/rpm/

3、安装

rpm -ivh *.rpm --force --nodeps

4、修改数据、日志目录

vi /etc/clickhouse-server/config.xml

5、创建相关目录

只需创建log文件目录即可

mkdir /bigdata/clickhouse

mkdir /bigdata/clickhouse/log

chown clickhouse:clickhouse /bigdata/clickhouse

chown clickhouse:clickhouse /bigdata/clickhouse/log

6、启动

/etc/init.d/clickhouse-server start

/etc/init.d/clickhouse-server stop

7、登录

由于9000与其它服务端口冲突，故tcp端口更改为9011

clickhouse-client --host localhost --port 9011

二、集群版

1、修改配置文件（三分片、单副本）

分别在三个节点都创建文件，红色字体每个节点配置不一样

vim /etc/metrika.xml

<yandex>
    <clickhouse_remote_servers>
        <cluster-01>
            <shard>
                <internal_replication>true</internal_replication>
                <replica>
                    <host>spbsjzy19</host>
                    <port>9011</port>
                    <user>rt</user>
                    <password>SPBsjzy@)@)</password>
                </replica>
            </shard>
            <shard>
                <internal_replication>true</internal_replication>
                <replica>
                    <host>spbsjzy20</host>
                    <port>9011</port>
                    <user>rt</user>
                    <password>SPBsjzy@)@)</password>
                </replica>
            </shard>
            <shard>
                <internal_replication>true</internal_replication>
                <replica>
                    <host>spbsjzy21</host>
                    <port>9011</port>
                    <user>rt</user>
                    <password>SPBsjzy@)@)</password>
                </replica>
            </shard>
        </cluster-01>
    </clickhouse_remote_servers>
    <zookeeper-servers>
        <node index="1">
            <host>spbsjzy33</host>
            <port>2181</port>
        </node>
        <node index="2">
            <host>spbsjzy34</host>
            <port>2181</port>
        </node>
        <node index="3">
            <host>spbsjzy35</host>
            <port>2181</port>
        </node>
    </zookeeper-servers>
     <macros>
      <cluster>cluster-01</cluster>
      <shard>01</shard>
      <replica>spbsjzy19</replica>
     <!-- <shard>02</shard>
      <replica>spbsjzy20</replica>
      <shard>03</shard>
      <replica>spbsjzy21</replica>
-->
    </macros>
    <clickhouse_compression>
        <case>
            <min_part_size>10000000000</min_part_size>
            <min_part_size_ratio>0.01</min_part_size_ratio>
            <method>lz4</method>
        </case>
    </clickhouse_compression>
</yandex>

注意：

1) <macros> </macros>中的shard名

三个节点配置相同则在任意一个节点插入数据，其他节点都会查询到；不相同的话复制表之间数据不会同步(插入分布表，数据会随机分配到某个节点的复制表；插入任意一个节点的复制表，分布表可以查询到，其他节点复制表无法查到)

2) <macros> </macros>中 <cluster>表示集群名称，<shard>表示分片编号，<replica>表示副本标识，这里使用了cluster{cluster}-{shard}-{replica}的表示方式

3) <internal_replication>

如果设置为true，则往本地表写入数据时，总是写入到完整健康的副本里，然后由表自身完成复制，这就要求本地表是能自我复制的（推荐）。如果设置为false，则写入数据时，是写入到所有副本中。这时，是无法保证一致性的

4) <user></user><password></password>

添加完后使用默认default用户也可以查询分布表。

相关报错信息：

ru.yandex.clickhouse.except.ClickHouseException: ClickHouse exception, code: 516, host: spbsjzy20, port: 8123; Code: 516, e.displayText() = DB::Exception: Received from spbsjzy19:9011. DB::Exception: default: Authentication failed: password is incorrect or there is no user with such name. (version 20.3.6.40 (official build))

2、修改配置

vim /etc/clickhouse-server/config.xml

把注释打开，否则节点之间无法访问，分布表也无法查询其它节点数据

3、建议

生产中建议一般采用复制表和分布表；先创建复制表而后创建分布表，复制表用来存储数据，分布表用来查询和写入

ReplacingMergeTree:

4、创建表

分别在三个节点创建复制表：

CREATE TABLE default.test (`eventdate` Date, `company` String, `deliveryno` String, `usercardtype` String, `id` String, `name` String, `mob` String, `orgcode` String, `creditcode` String, `taxregno` String, `type` String, `cardid` String, `staffna` String, `staffmob` String, `staffaddress` String, `checkdate` String, `method` String, `address` String, `utcdate` DateTime DEFAULT now()) ENGINE = ReplicatedMergeTree('/clickhouse/tables/{shard}/test', '{replica}', eventdate, (eventdate, company, deliveryno, id, mob, name), 8192)

在一个节点创建分布表：

CREATE TABLE default.test_all (`eventdate` Date, `company` String, `deliveryno` String, `usercardtype` String, `id` String, `name` String, `mob` String, `orgcode` String, `creditcode` String, `taxregno` String, `type` String, `cardid` String, `staffna` String, `staffmob` String, `staffaddress` String, `checkdate` String, `method` String, `address` String, `utcdate` DateTime DEFAULT now()) ENGINE = Distributed('{cluster}', 'default', 'test', rand())

5、插入表

insert into test(eventdate,company,deliveryno,id) VALUES('2020-12-05','ff','vv','xx');
insert into test_all(eventdate,company,deliveryno,id) VALUES('2020-12-06','xx','vv','xx');

6、数据分布测试

（1）三张复制表分别执行

insert into test(eventdate,company,deliveryno,id) VALUES('2020-12-07','ff','vv','xx');
insert into test(eventdate,company,deliveryno,id) VALUES('2020-12-08','ff','vv','xx');
insert into test(eventdate,company,deliveryno,id) VALUES('2020-12-09','ff','vv','xx');

查询复制表：

查询分布表

（2）插入分布表（随机分布）

insert into test_all(eventdate,company,deliveryno,id) VALUES('2020-12-07','xx','vv','xx');
insert into test_all(eventdate,company,deliveryno,id) VALUES('2020-12-08','xx','vv','xx');

查询复制表

查询分布表

7、节点宕机测试

加载全部内容