Druid.io系列（八）：部署

介绍

前面几个章节对Druid的整体架构做了简单的说明，本文主要描述如何部署Druid的环境

Imply提供了一套完整的部署方式，包括依赖库，Druid，图形化的数据展示页面，SQL查询组件等。本文将基于Imply套件进行说明

单机部署

依赖

• Java 8 or better
• Node.js 4.5.x or better
• Linux, Mac OS X, or other Unix-like OS (Windows is not supported)
• At least 4GB of RAM

下载与安装

1. 从https://imply.io/get-started 下载最新版本安装包
2. tar -xzf imply-2.5.14.tar.gz
3. cd imply-2.5.14

目录说明如下： 
- bin/ - run scripts for included software. 
- conf/ - template configurations for a clustered setup. 
- conf-quickstart/* - configurations for the single-machine quickstart. 
- dist/ - all included software. 
- quickstart/ - files related to the single-machine quickstart.

启动服务

bin/supervise -c conf/supervise/quickstart.conf

安装验证

导入测试数据

安装包中包含一些测试的数据,可以通过执行预先定义好的数据说明文件进行导入

bin/post-index-task --file quickstart/wikiticker-index.json

可视化控制台

• overlord 控制页面：http://localhost:8090/console.html.
• druid集群页面：http://localhost:8081
• 数据可视化页面：http://localhost:9095

1. HTTP POST数据查询

命令：curl -L -H’Content-Type: application/json’ -XPOST –data-binary @quickstart/wikiticker-top-pages.json http://localhost:8082/druid/v2?pretty 
结果：

[ {   "timestamp" : "2016-06-27T00:00:11.080Z",   "result" : [ {     "edits" : 29,     "page" : "Copa América Centenario"   }, {     "edits" : 16,     "page" : "User:Cyde/List of candidates for speedy deletion/Subpage"   },   ..........   {     "edits" : 8,     "page" : "World Deaf Championships"   } ] } ]

集群部署

集群配置的规划需要根据需求来定制，下面以一个开发环境机器搭建为例,描述如何搭建一个有HA特性的Druid集群.

集群部署有以下几点需要说明 
1. 为了保证HA,主节点部署两台 
2. 管理节点与查询节点可以考虑多核大内存的机器

部署规划

角色	机器	配置	集群角色
主节点	10.5.24.137	8C16G	Coordinator,Overlord
主节点	10.5.24.138	8C16G	Coordinator,Overlord
数据节点，查询节点	10.5.24.139	8C16G	Historical, MiddleManager, Tranquility,Broker,Pivot Web
数据节点，查询节点	10.5.24.140	8C16G	Historical, MiddleManager, Tranquility,（数据节点，查询节点）Broker

部署步骤

公共配置

编辑conf/druid/_common/common.runtime.properties 文件内容 

#
# Extensions
#

druid.extensions.directory=dist/druid/extensions
druid.extensions.hadoopDependenciesDir=dist/druid/hadoop-dependencies
druid.extensions.loadList=["druid-parser-route","druid-lookups-cached-global","druid-histogram","druid-datasketches","mysql-metadata-storage","druid-hdfs-storage","druid-kafka-indexing-service"]

#
# Logging
#

# Log all runtime properties on startup. Disable to avoid logging properties on startup:
druid.startup.logging.logProperties=true

#
# Zookeeper
#
druid.zk.service.host=master251
druid.zk.paths.base=/druid

# For MySQL:
druid.metadata.storage.type=mysql
druid.metadata.storage.connector.connectURI=jdbc:mysql://master251:3306/druid
druid.metadata.storage.connector.user=druid
druid.metadata.storage.connector.password=druid

#
# Deep storage
#
# For local disk (only viable in a cluster if this is a network mount):
# For HDFS:
druid.storage.type=hdfs
druid.storage.storageDirectory=hdfs://master251:9000/druid/segments

主节点配置

1. 创建配置文件：cp conf/supervise/master-no-zk.conf conf/supervise/master.conf
2. 编辑master.conf 内容如下：

:verify bin/verify-java
:verify bin/verify-version-check

coordinator bin/run-druid coordinator conf
!p80 overlord bin/run-druid overlord conf

目前的版本中，mysql-metadata-storage没有包含在默认的安装包中，如果使用mysql存储元数据，需要单独安装下对应的扩展，是用下列命令在两个master节点上对需要用到的扩展进行安装：

root@druid imply-2.5.14# java -classpath "dist/druid/lib/*"  -Ddruid.extensions.directory="dist/druid/extensions" io.druid.cli.Main tools pull-deps  -c io.druid.extensions:mysql-metadata-storage:0.12.0 -c io.druid.extensions.contrib:druid-rabbitmq:0.12.0 -h org.apache.hadoop:hadoop-client:2.6.0

==默认mysql-metadata-storage带的mysql驱动是针对Mysql 5.1的，如果使用Mysql的版本是5.5 或是其他版本，可能会出现”Communications link failure”的错误，此时需要更新Mysql的驱动。==

在10.5.24.137/138上启动master相关服务：nohup bin/supervise -c conf/supervise/master.conf > master.log &

数据节点与查询节点配置

1. 安装NodeJS:apt-get install nodejs
2. 创建配置文件：vim conf/supervise/data-with-query.conf
3. 编辑data-with-query.conf 内容如下：

:verify bin/verify-java
:verify bin/verify-node
:verify bin/verify-version-check

broker bin/run-druid broker conf
imply-ui bin/run-imply-ui conf


historical bin/run-druid historical conf
middleManager bin/run-druid middleManager conf

# Uncomment to use Tranquility Server
#!p95 tranquility-server bin/tranquility server -configFile conf/tranquility/server.json

# Uncomment to use Tranquility Kafka
#!p95 tranquility-kafka bin/tranquility kafka -configFile conf/tranquility/kafka.json

对于集群模式，pivot的配置文件必须调整为mysql,sqllite会导致无法查看datasource,修改conf/pivot/config.xml文件

settingsLocation:
  location: mysql
  uri: 'mysql://root:123456@10.5.24.151:3306/druid'
  table: 'pivot_state'
  initialSettings:
  clusters:
    - name: druid
      type: druid
      host: localhost:8082

在10.5.24.139/140两台机器上分别执行:nohup bin/supervise -c conf/supervise/data-with-query.conf > data-with-query.log &

验证

可视化控制台

• overlord 控制页面：http://10.5.24.138:8090/console.html.
• druid集群页面：http://10.5.24.138:8081
• 数据可视化页面：http://10.5.24.139:9095