Hadoop学习笔记(1)-基本概念

时间:2022-11-12 04:09:55
        Most of us would be specialized in one or a few specific area/layer as an engineer. Technically it is understandable because not a single person can master everything, but I strongly recommend you to try to have some big picture of the whole process. Whenever you have some issues or something for you to work, try to ask your self "Where is the current issue located in the whole picture ?".
       You don't have to try rote memorization for these whole process, but if you keep trying to map your current issues to specific parts within the whole picture, these whole sequence would automatically pops up in your memory. I say so often to my kid or my juniors "Don't try to memorize it, just repeat it until you automatically recall" (They are looking at me seeminly saying "What is the difference ?" -:) I just smile -:).


1、什么是hadoop

(1)HADOOP是apache旗下的一套开源软件平台
(2)HADOOP提供的功能:利用服务器集群,根据用户的自定义业务逻辑,对海量数据进行分布式处理
(3)HADOOP的核心组件有
A.HDFS(分布式文件系统)
B.YARN(运算资源调度系统)
C.MAPREDUCE(分布式运算编程框架)
(4)广义上来说,HADOOP通常是指一个更广泛的概念——HADOOP生态圈
2、HADOOP生态圈以及各组成部分的简介


各组件简介
Hadoop学习笔记(1)-基本概念
重点组件:
HDFS:分布式文件系统
MAPREDUCE:分布式运算程序开发框架
HIVE:基于大数据技术(文件系统+运算框架)的SQL数据仓库工具
HBASE:基于HADOOP的分布式海量数据库
ZOOKEEPER:分布式协调服务基础组件
Mahout:基于mapreduce/spark/flink等分布式运算框架的机器学习算法库
Oozie:工作流调度框架
Sqoop:数据导入导出工具
Flume:日志数据采集框架

3、项目基本架构

Hadoop学习笔记(1)-基本概念

1)数据采集:定制开发采集程序(shell脚步),或使用开源框架FLUME
2)数据预处理:定制开发mapreduce程序运行于hadoop集群
3)数据仓库技术:基于hadoop之上的Hive
4)数据导出:基于hadoop的sqoop数据导入导出工具
5)数据可视化:定制开发web程序或使用kettle等产品
6)整个过程的流程调度:hadoop生态圈中的oozie工具或其他类似开源产品