clickhouse 22.8.12.45 详细部署

1.包下载

下载地址：https://packages.clickhouse.com/rpm/stable/
下载以下几个包：
clickhouse-client-22.8.12.45.x86_64.rpm
clickhouse-common-static-22.8.12.45.x86_64.rpm
clickhouse-common-static-dbg-22.8.12.45.x86_64.rpm
clickhouse-keeper-22.8.12.45.x86_64.rpm【可以不下载，本文使用zk，后续更新迁移至 ckeeper】
clickhouse-server-22.8.12.45.x86_64.rpm

2.安装

rpm -ivh clickhouse-common-static-dbg-22.8.12.45.x86_64.rpm
rpm -ivh clickhouse-common-static-22.8.12.45.x86_64.rpm
rpm -ivh clickhouse-server-22.8.12.45.x86_64.rpm -- 按回车先跳过，后面再设置
rpm -ivh clickhouse-client-22.8.12.45.x86_64.rpm

3.目录说明

/etc/clickhouse-server	服务端的配置文件目录
/etc/clickhouse-client	客户端配置
/var/lib/clickhouse	默认的数据存储目录
/var/log/clickhouse-server	默认保存日志的目录

4.修改配置文件 config.xml

1.修改时区：<timezone>Asia/Shanghai</timezone>

2.开启外部访问：<listen_host>0.0.0.0</listen_host>

3.修改数据存放目录：<path>/var/lib/clickhouse/</path>

4.修改日志存放目录：
<log>/var/log/clickhouse-server/clickhouse-server.log</log>
<errorlog>/var/log/clickhouse-server/clickhouse-server.err.log</errorlog>

5.配置文件句柄数量：<max_open_files>262144</max_open_files>

6.http默认端口8123：<http_port>8123</http_port>

7.tcp默认端口9000：<tcp_port>9000</tcp_port>

8.zk配置：
    <zookeeper>
        <node>
            <host>hdfs1</host>
            <port>2181</port>
        </node>
        <node>
            <host>hdfs2</host>
            <port>2181</port>
        </node>
        <node>
            <host>hdfs3</host>
            <port>2181</port>
        </node>
    </zookeeper>

9.集群配置
（3分片1副本）
    <remote_servers>
        <CK_3shard_1rep>
            <shard>
                <internal_replication>true</internal_replication>
                <replica>
                    <host>hdfs1</host>
                    <port>9000</port>
               </replica>
            </shard>
            <shard>
                <internal_replication>true</internal_replication>
                <replica>
                    <host>hdfs2</host>
                    <port>9000</port>
                </replica>
            </shard>
            <shard>
                <internal_replication>true</internal_replication>
                <replica>
                    <host>hdfs3</host>
                    <port>9000</port>
              </replica>
            </shard>
        </CK_3shard_1rep>
    </remote_servers>

10.配置宏，三台机器不一致
hdfs1：
<macros>
	<shard>01</shard>
    <replica>hdfs1</replica>
</macros>
hdfs2：
<macros>
	<shard>02</shard>
    <replica>hdfs2</replica>
</macros>
hdfs3：
<macros>
	<shard>03</shard>
    <replica>hdfs3</replica>
</macros>

5.修改密码

本文使用 sha256加密，官方推荐
密码为：123456

[root@hdfs2 clickhouse-server]# echo -n "123456" | sha256sum | tr -d '-'
8d969eef6ecad3c29a3a629280e686cf0c3f5d5a86aff3ca12020c923adc6c92

将以上密码修改到 /etc/clickhouse-server/users.d/default-password.xml
先备份下：cp /etc/clickhouse-server/users.d/default-password.xml /etc/clickhouse-server/users.d/default-password.xml_bak

<clickhouse>
    <users>
        <default>
            <password remove='1' />
            <password_sha256_hex>8d969eef6ecad3c29a3a629280e686cf0c3f5d5a86aff3ca12020c923adc6c92</password_sha256_hex>
        </default>
    </users>
</clickhouse>

6.启动并连接

# 启动
systemctl start clickhouse-server
# 状态
systemctl status clickhouse-server
# 连接
clickhouse-client -n -m  -h hdfs1 --port 9000 -udefault --password 123456
# 查看集群分片（3分片1副本）
hdfs1 :) select * from system.clusters;

SELECT *
FROM system.clusters

Query id: a34401a0-ba49-486b-88ad-6718521d0e81

┌─cluster────────┬─shard_num─┬─shard_weight─┬─replica_num─┬─host_name─┬─host_address────┬─port─┬─is_local─┬─user────┬─default_database─┬─errors_count─┬─slowdowns_count─┬─estimated_recovery_time─┐
│ CK_3shard_1rep │         1 │            1 │           1 │ hdfs1     │ 192.168.136.221 │ 9000 │        1 │ default │                  │            0 │               0 │   0 │
│ CK_3shard_1rep │         2 │            1 │           1 │ hdfs2     │ 192.168.136.222 │ 9000 │        0 │ default │                  │            0 │               0 │   0 │
│ CK_3shard_1rep │         3 │            1 │           1 │ hdfs3     │ 192.168.136.223 │ 9000 │        0 │ default │                  │            0 │               0 │   0 │
└────────────────┴───────────┴──────────────┴─────────────┴───────────┴─────────────────┴──────┴──────────┴─────────┴──────────────────┴──────────────┴─────────────────┴─────────────────────────┘

3 rows in set. Elapsed: 0.001 sec.

7.创建本地表&分布式表

# 本地表
CREATE TABLE default.test001_local on cluster CK_3shard_1rep(    
  _pos UInt8,
  _user UInt64,
  _exposure_num UInt64,
  _type String,
  datekey String
)
ENGINE = ReplicatedReplacingMergeTree() PARTITION BY datekey ORDER BY (_pos)

# 分布式表
CREATE TABLE default.test001_distributed on cluster CK_3shard_1rep(    
  _pos UInt8,
  _user UInt64,
  _exposure_num UInt64,
  _type String,
  datekey String
)
ENGINE = Distributed('CK_3shard_1rep', 'default', 'test001_local', rand())

8.附加配置说明

<listen_host>::</listen_host>	允许IP4和IP6源主机远程访问
<listen_host>0.0.0.0</listen_host>	仅允许IP4主机远程访问
<listen_host>::1</listen_host> <listen_host>127.0.0.1</listen_host>	仅允许本地访问

到这边简单的clickhouse集群就部署完成，进行了简单的配置，以上配置仅供参考！！！
原理、特性、架构，还在慢慢熟悉，让我们一起学起来吧！！！