Install MapR Cluster
Quick and dirty way to install MapR cluster within few mins with just few commands if you don't want to script this process or use GUI installer.
In this post its assumed you have a system with Red-Hat/CentOS 6.x already then below steps required to get MapR Hadoop Cluster up and running.
1) yum install wget -y
2) wget http://download.fedoraproject.org/pub/epel/6/x86_64/epel-release-6-8.noarch.rpm
3) rpm -Uvh epel-release-6*.rpm
4) rpm --import http://package.mapr.com/releases/pub/maprgpg.key
5) groupadd -g 5000 mapr ; useradd -g 5000 -u 5000 mapr
6) vi /etc/yum.repos.d/maprtech.repo ---- Add below address
[maprtech]
name=MapR Technologies
baseurl=http://package.mapr.com/releases/v5.2.2/redhat/
enabled=1
gpgcheck=0
protect=1
[centos]
name=centos, Inc.
baseurl=http://mirror.centos.org/centos/6.9/os/x86_64/
enabled=1
gpgcheck=0
7) Now creating list of disk to be given to MFS, First disk is OS disk hence came up with below script to get list of disks.
( Below is example to get disk list on aws instance excluding OS disk )
fdisk -l | grep dev | awk '{print $2}' | cut -d: -f1 | tail -n +2 >> /tmp/disks
8) Install all required services with Java.
yum install mapr-cldb mapr-core mapr-fileserver mapr-zookeeper mapr-zk-internal mapr-webserver java-devel -y
9) Add "/usr/lib/jvm/jre-1.6.0-openjdk.x86_64" to below env file as "JAVA_HOME"
vi /opt/mapr/conf/env.sh
10) /opt/mapr/server/configure.sh -C Nodes -Z Nodes -N Cluster-Name
11) /opt/mapr/server/disksetup -F /tmp/disks
12) Start ZK service and check qstatus if it has leader.
service mapr-zookeeper start
service mapr-zookeeper qstatus
13) Start warden
service mapr-warden start
14) Check MCS to Apply License and monitor the cluster.
connect browser to https://{webserver host name}:8443
No comments:
Post a Comment