Before we start with the installations, it is important to make sure that the host resolution is configured and working properly.
Choose any appropriate hostnames the user wants for his or her Linux machines. For example, the hostnames could be master1.cluster.com or rt1.cyrus.com or host1.example.com. The important thing is that the hostnames must resolve.
This resolution can be done using a DNS server or by configuring the/etc/hosts file on each node we use for our cluster setup.
The following steps will show you how to set up the resolution in the/etc/hosts file.
Connect to the Linux machine and change the hostname to
master1.cyrus.comin the file as follows:
Edit the
/etc/hostsfile as follows:
Make sure the resolution returns an IP address:
# getent hosts master1.cyrus.comThe other preferred method is to set up the DNS resolution so that we do not have to populate the
hostsfile on each node. In the example resolution shown here, the user can see that the DNS server is configured to answer the domaincyrus.com:# nslookup master1.cyrus.com Server: 10.0.0.2 Address: 10.0.0.2#53 Non-authoritative answer: Name: master1.cyrus.com Address: 10.0.0.104
Each Linux host has a resolver library that helps it resolve any hostname that is asked for. It contacts the DNS server, and if it is not found there, it contacts the hosts file. Users who are not Linux administrators can simply use the hosts files as a workaround to set up a Hadoop cluster. There are many resources available online that could help you to set up a DNS quickly if needed.
Once the resolution is in place, we will start with the installation of Hadoop on a single-node and then progress to multiple nodes.
