Clickhouse Centos74 rpm 安装

Clickhouse Centos74 rpm 安装

 

安装环境

华为云主机双cpu双核心，4G内存，centos7.4

192.168.1.223 ecs-app-0001.novalocal ecs-app-0001

192.168.1.240 ecs-app-0002.novalocal ecs-app-0002

192.168.1.197 ecs-app-0003.novalocal ecs-app-0003

 

详细命令请参考:

https://github.com/super-sponge/clickhose_deploy

 

1. ansible 安装及配置

yum install -y ansible

1.1 去掉 /etc/ansible/ansible.cfg host_key_checking 前面的注释

host_key_checking = False

1.2 编辑ansible host文件 (root 密码根据实际情况填写)

[clickhouse]

192.168.1.223 ansible_ssh_user=root ansible_ssh_pass=****

192.168.1.240 ansible_ssh_user=root ansible_ssh_pass=****

192.168.1.197 ansible_ssh_user=root ansible_ssh_pass=****

 

2. 安装clickhose

2.1 安装clickhouse

ansible clickhouse -m copy -a 'src=Altinity_clickhouse.repo dest=/etc/yum.repos.d'

ansible all -m yum -a "name=clickhouse-server state=present"

ansible all -m yum -a "name=clickhouse-client state=present"

 

备注：

也可以手工在每台主机上执行

curl -s https://packagecloud.io/install/repositories/Altinity/clickhouse/script.rpm.sh | bash

yum install -y clickhouse-server

yum install -y clickhouse-client

 

2.2 配置

配置config.xml 文件

vi /etc/clickhouse-server/config.xml

/data/clickhouse/

/data/clickhouse/tmp/

ecs-app-0001.novalocal

0.0.0.0

 

备注：

interserver_http_host 需要配置为实际的主机名

ansible all -m copy -a 'src=config.xml dest=/etc/clickhouse-server mode=0644'

进入每个主机更改interserver_http_host

 

配置分布式环境users.xml 和metrika.xml(实际内容请参考附件)

ansible all -m copy -a 'src=users.xml dest=/etc/clickhouse-server mode=0644'

ansible all -m copy -a 'src=metrika.xml dest=/etc/metrika.xml mode=0644'

 

备注:

metrika.xml 中（参考网上的资料，我对每个主机设置为不同的值，这个上面批量拷贝后，手工更改）

ck1

3. 安装 zookeeper 以及jdk

请参考附件

 

4. 启动与停止

ansible clickhouse -m service -a "name=clickhouse-server state=started"

ansible clickhouse -m service -a "name=clickhouse-server state=stopped"

5. 测试

表创建

./create.sh

装在数据

./load.sh

cat create.sh

#!/bin/bash

 

base_dir=$(cd $(dirname $0); pwd)

 

 

for i in 1 2 3

do

clickhouse-client --host=ecs-app-000${i}.novalocal --password=6lYaUiFi -n < $base_dir/create.sql

done

 

load 数据

load.sh

#!/bin/bash

 

base_dir=$(cd $(dirname $0); pwd)

 

cd $base_dir/dss

 

./dbgen -s 10 -T a

 

sed -i 's/$/"1999-01-01"/' customer.tbl

cat customer.tbl | clickhouse-client --host=ecs-app-0001.novalocal --password=6lYaUiFi --query "INSERT INTO customerd FORMAT CSV"

cat lineorder.tbl | clickhouse-client --host=ecs-app-0001.novalocal --password=6lYaUiFi --query "INSERT INTO lineorderd FORMAT CSV"

cat part.tbl | clickhouse-client --host=ecs-app-0001.novalocal --password=6lYaUiFi --query "INSERT INTO partd FORMAT CSV"

cat supplier.tbl | clickhouse-client --host=ecs-app-0001.novalocal --password=6lYaUiFi --query "INSERT INTO supplierd FORMAT CSV"

 

 

ntp 时间同步配置

ansible all -m yum -a "name=chrony state=present"

 

 

附件：

1. 安装zookeeper 和 jdk

jdk 安装 （本次安装到全局环境）

ansible clickhouse -m file -a 'path=/usr/java state=directory recurse=yes mode=0755'

ansible clickhouse -m unarchive -a "src=jdk-8u191-linux-x64.tar.gz dest=/usr/java"

 

vi jdk.sh

JAVA_HOME=/usr/java/jdk1.8.0_191

CLASSPATH=$JAVA_HOME/lib/

PATH=$PATH:$JAVA_HOME/bin

export PATH JAVA_HOME CLASSPATH

 

ansible clickhouse -m copy -a "src=jdk.sh dest=/etc/profile.d"

 

zookeeper 安装

ansible clickhouse -m group -a 'name=hadoop state=present'

ansible clickhouse -m user -a 'name=zookeeper shell=/bin/bash group=hadoop state=present'

ansible clickhouse -m file -a 'path=/home/zookeeper/app state=directory recurse=yes owner=zookeeper group=hadoop mode=0744'

ansible clickhouse -m unarchive -a "src=zookeeper-3.4.13.tar.gz dest=/home/zookeeper/app owner=zookeeper group=hadoop"

 

ansible clickhouse -m file -a 'path=/data/zookeeper/data state=directory recurse=yes owner=zookeeper group=hadoop mode=0744'

ansible clickhouse -m file -a 'path=/data/zookeeper/datalog state=directory recurse=yes owner=zookeeper group=hadoop mode=0744'

 

vi zoo.cfg

 

tickTime=2000

initLimit=10

syncLimit=5

dataDir=/data/zookeeper/data

dataLogDir=/data/zookeeper/datalog

clientPort=2181

server.0=ecs-app-0001.novalocal:2888:3888

server.1=ecs-app-0002.novalocal:2888:3888

server.2=ecs-app-0003.novalocal:2888:3888

 

ansible clickhouse -m copy -a 'src=zoo.cfg dest=/home/zookeeper/app/zookeeper-3.4.13/conf owner=zookeeper group=hadoop mode=0644'

 

vi zk.sh

#!/bin/bash

 

for id in 0 1 2

do

echo $id > myid

ansible clickhouse[$id] --become-user zookeeper -b -m copy -a "src=myid dest=/data/zookeeper/data owner=zookeeper group=hadoop"

done

 

启动(下面启动没有成功，直接登录每台主机启动)

ansible clickhouse --become-user zookeeper -b -m command -a "/home/zookeeper/app/zookeeper-3.4.13/bin/zkServer.sh start"

停止

ansible clickhouse --become-user zookeeper -b -m command -a "/home/zookeeper/app/zookeeper-3.4.13/bin/zkServer.sh stop"

验证

./zkCli.sh

create /zk_test my_data

 

2. metrika.xml

 

 

false

ecs-app-0001.novalocal

9000

default

6lYaUiFi

 

false

ecs-app-0002.novalocal

9000

default

6lYaUiFi

 

false

ecs-app-0003.novalocal

9000

default

6lYaUiFi

 

 

ck1

 

::/0

 

ecs-app-0001.novalocal

2181

ecs-app-0002.novalocal

2181

ecs-app-0003.novalocal

2181

 

10000000000

0.01

lz4

 

 

3. users.xml

10000000000

0

random

 

10000000000

0

random

1

 

3600

0

0

0

0

0

 

967f3bf355dddfabfca1c9f5cab39352b2ec1cd0b05f9e1e6b8f629705fe7d6e

::/0

0.0.0.0

default

default

 

967f3bf355dddfabfca1c9f5cab39352b2ec1cd0b05f9e1e6b8f629705fe7d6e

::/0

0.0.0.0

readonly

default

 

4. create.sql

drop table if exists lineorder;

drop table if exists customer;

drop table if exists part;

drop table if exists supplier;

drop table if exists lineorderd;

drop table if exists customerd;

drop table if exists partd;

drop table if exists supplierd;

 

 

CREATE TABLE lineorder (

LO_ORDERKEY UInt32,

LO_LINENUMBER UInt8,

LO_CUSTKEY UInt32,

LO_PARTKEY UInt32,

LO_SUPPKEY UInt32,

LO_ORDERDATE Date,

LO_ORDERPRIORITY String,

LO_SHIPPRIORITY UInt8,

LO_QUANTITY UInt8,

LO_EXTENDEDPRICE UInt32,

LO_ORDTOTALPRICE UInt32,

LO_DISCOUNT UInt8,

LO_REVENUE UInt32,

LO_SUPPLYCOST UInt32,

LO_TAX UInt8,

LO_COMMITDATE Date,

LO_SHIPMODE String

)Engine=MergeTree(LO_ORDERDATE,(LO_ORDERKEY,LO_LINENUMBER),8192);

 

CREATE TABLE customer (

C_CUSTKEY UInt32,

C_NAME String,

C_ADDRESS String,

C_CITY String,

C_NATION String,

C_REGION String,

C_PHONE String,

C_MKTSEGMENT String,

C_FAKEDATE Date

)Engine=MergeTree(C_FAKEDATE,(C_CUSTKEY),8192);

 

CREATE TABLE part (

P_PARTKEY UInt32,

P_NAME String,

P_MFGR String,

P_CATEGORY String,

P_BRAND String,

P_COLOR String,

P_TYPE String,

P_SIZE UInt8,

P_CONTAINER String,

P_FAKEDATE Date

)Engine=MergeTree(P_FAKEDATE,(P_PARTKEY),8192);

 

CREATE TABLE supplier (

S_SUPPKEY UInt32,

S_NAME String,

S_ADDRESS String,

S_CITY String,

S_NATION String,

S_REGION String,

S_PHONE String,

S_FAKEDATE Date

)Engine=MergeTree(S_FAKEDATE,(S_SUPPKEY),8192);

 

 

CREATE TABLE lineorderd AS lineorder ENGINE = Distributed(bip_ck_cluster, default, lineorder, rand());

CREATE TABLE customerd AS customer ENGINE = Distributed(bip_ck_cluster, default, customer, rand());

CREATE TABLE partd AS part ENGINE = Distributed(bip_ck_cluster, default, part, rand());

CREATE TABLE supplierd AS supplier ENGINE = Distributed(bip_ck_cluster, default, supplier, rand());