【1】搭建HA高可用hadoop-2.3(規劃+環境準備)java
【2】搭建HA高可用hadoop-2.3(安裝zookeeper)node
【3】搭建HA高可用hadoop-2.3(部署配置hadoop--cdh5.1.0)linux
【4】搭建HA高可用hadoop-2.3(部署配置HBase)centos
搭建HA高可用hadoop-2.3bash
前言:cookie
做爲安裝Hadoop的第一步,就是根據實際狀況選擇合適的Hadoop版本,此次我所用的CDH5.1.0,基於Hadoop2.3版本。那麼什麼是CDH呢,下面科普一下。app
Cloudera是一家提供Hadoop支持、諮詢和管理工具的公司,在Hadoop生態圈有着舉足輕重的地位,它的拳頭產品就是著名的Cloudera's Distribution for Hadoop,簡稱CDH。該軟件同咱們熟知的Apache Hadoop同樣,都是徹底開源,基於Apache軟件許可證,免費爲我的和商業使用。Coudera從一個穩定的Apache Hadoop版本開始,接二連三的發佈新版本併爲舊版本打上補丁,爲各類不一樣的生產環境提供安裝文件,在Cloudera的團隊中有許多Apache Hadoop的代碼貢獻者,因此Cloudera的公司實力毋庸置疑。ssh
通常用戶安裝Hadoop時,不單單隻安裝HDFS、MapReduce,還會根據須要安裝Hive、HBase、Spark等。Cloudera將這些相關的項目都集成在一個CDH版本里面,目前CDH包括Hadoop、HBase、Hive、Pig、Sqood、Zooksspe、Spark、Flume、Oozie、Mahout等等,幾乎覆蓋了Hadoop生態圈,這樣作的好處是保證了組件之間的兼容性,由於各個項目之間也存在徹底獨立的版本,其各個版本與Hadoop之間必然會存在兼容性的問題,若是選擇CDH,那麼同一個CDH版本內的各個組建將徹底不存在兼容性問題。因此初學者使用CDH來搭建Hadoop是一個很好的選擇。tcp
規劃:ide
機器:5臺 OS:centos6.5
主機名 |
IP | HDFS | Yarn | HBase |
master1 | 10.64.8.1 | Namenode1 | ResourceManager | HMaster |
master1 | 10.64.8.2 | Namenode2 | ||
slave1 | 10.64.8.3 | Datanode1 | NodeManager | HRegionServer |
slave2 | 10.64.8.4 | Datanode2 | NodeManager | HRegionServer |
slave3 | 10.64.8.5 | Datanode3 | NodeManager | HRegionServer |
準備工做:
(1)關閉selinux和防火牆
master一、master二、slave一、slave二、slave3
#setenforce 0&& service iptables stop &&chkconfig iptables off
(2)修改主機名
master1
#hostname master1 && echo master1 >/etc/hostname
master2
#hostname master1 && echo master1 >/etc/hostname
slave1
#hostname slave1 && echo slave1 >/etc/hostname
slave2
#hostname slave2 && echo slave2 >/etc/hostname
slave3
#hostname slave3 && echo slave3 >/etc/hostname
master一、master二、slave一、slave二、slave3
#cat >> /etc/hosts << EOF 10.64.8.1 master1 10.64.8.2 master2 10.64.8.3 slave1 10.64.8.4 slave2 10.64.8.5 slave3 EOF
(3)主從節點ssh互信
生產中使用hadoop用戶來運行,root用戶會有風險,初學者可使用root用戶,避免權限問題的困擾。
master一、master2
#ssh-keygen -t rsa #ssh-copy-id -i ~/.ssh/id_rsa.pub root@slave1 #ssh-copy-id -i ~/.ssh/id_rsa.pub root@slave2 #ssh-copy-id -i ~/.ssh/id_rsa.pub root@slave3
(4)安裝jdk
master一、master二、slave一、slave二、slave3
#rpm -e --nodeps java-1.6.0-openjdk ---刪除openjdk #yum install jdk1.8.0_60.x86_64
(5)格式化硬盤
master一、master二、slave一、slave二、slave3
#mkfs.ext4 /dev/sdb #mkdir /data #mount /dev/sdb /data #echo "mount /dev/sdb /data" >>/etc/rc.local
(6)系統參數調優
master一、master二、slave一、slave二、slave3
cat >>/etc/sysctl.conf <<EOF net.ipv4.tcp_fin_timeout = 30 net.ipv4.tcp_keepalive_time = 1200 net.ipv4.tcp_syncookies = 1 net.ipv4.tcp_tw_reuse = 1 net.ipv4.tcp_tw_recycle = 1 net.ipv4.ip_local_port_range = 1024 65000 net.ipv4.tcp_max_syn_baklog = 8192 net.ipv4.tcp_max_tw_bukets = 5000 EOF
cat >> /etc/security/limits.conf <<EOF * soft nofile 65535* hard nofile 65535 EOF
重啓
(7)下載須要的安裝包
#cd /opt #wget hadoop-2.3.0-cdh5.1.0.tar.gz
附CDH5.1.0下載連接
hadoop-2.3.0-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/hadoop-2.3.0-cdh5.1.0.tar.gz
zookeeper-3.4.5-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/zookeeper-3.4.5-cdh5.1.0.tar.gz
hive-0.12.0-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/hive-0.12.0-cdh5.1.0.tar.gz
hbase-solr-1.5-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/hbase-solr-1.5-cdh5.1.0.tar.gz
hbase-0.98.1-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/hbase-0.98.1-cdh5.1.0.tar.gz
spark-1.0.0-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/spark-1.0.0-cdh5.1.0.tar.gz
flume-ng-1.5.0-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/flume-ng-1.5.0-cdh5.1.0.tar.gz
solr-4.4.0-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/solr-4.4.0-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/mahout-0.9-cdh5.1.0.tar.gz
hue-3.6.0-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/hue-3.6.0-cdh5.1.0.tar.gz
oozie-4.0.0-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/oozie-4.0.0-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/whirr-0.9.0-cdh5.1.0.tar.gz
pig-0.12.0-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/pig-0.12.0-cdh5.1.0.tar.gz
search-1.0.0-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/search-1.0.0-cdh5.1.0.tar.gz
parquet-1.2.5-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/parquet-1.2.5-cdh5.1.0.tar.gz
parquet-format-1.0.0-cdh5.1.0.tar.gz
http://archive.cloudera.com/cdh5/cdh/5/parquet-format-1.0.0-cdh5.1.0.tar.gz
【2】搭建HA高可用hadoop-2.3(安裝zookeeper)