Cassandra Replication - HA

Introduction

Cassandra stores replicas on multiple nodes to ensure reliability and fault tolerance. A replication strategy determines the nodes where replicas are placed. The total number of replicas across the cluster is referred to as the replication factor.

A replication factor of 1 means that there is only one copy of each row in the cluster. A replication factor of 2 means two copies of each row, where each copy is on a different node. As a general rule, the replication factor should not exceed the number of nodes in the cluster. However, you can increase the replication factor and then add the desired number of nodes.

[Image unavailable in the original archive]

Prerequisites

To follow this tutorial :

• You need atleast two nodes.

• You need to install cassandra in all the nodes.

Step 1 - Create a Keyspace in cql

• login into cqlsh

    cqlsh `ipaddress1`
  

• Creating a keyspace

    CREATE KEYSPACE IF NOT EXISTS `keyspacename` WITH REPLICATION = { 'class' : 'NetworkTopologyStrategy', 'dc1' : 2 , 'dc2' : 2 };
  

• Two replication strategies of class are available:

• SimpleStrategy Use only for a single data center and one rack. If you ever intend more than one data center, use the NetworkTopologyStrategy.

• NetworkTopologyStrategy Highly recommended for most deployments because it is much easier to expand to multiple data centers when required by future expansion.

• Here, we are testing in multiple datacenter so we used NetworkTopologyStrategy class.

Step 2 - Configure cassandra.yaml in dc1

• Its assumed that you have installed cassandra in the two nodes (dc1, dc2).

• Once you installed cassandra in your machine, then you need to change the cassandra yaml file in (dc1).

• Open the file /etc/cassandra/cassandra.yaml and change the following settings

    listen_address : `ipaddress`
    rpc_address    : `ipaddress`
    endpoint_snitch: GossipingPropertyFileSnitch
  

Note listen_address and rpc_address default is set localhost you need to change it to the private or public ipaddress (eg: 192.168.1.249).

endpoint_snitch by default is set for SimpleSnitch which works only for SimpleStrategy so you can changed the snitch to GossipingPropertyFileSnitch

• As we have used GossipingPropertyFileSnitch we need to change the file /etc/cassandra/cassandra-rackdc.properties with the datacenter and rack information.

• cassandra-rackdc.properties it used to tell we are using GossipingPropertyFileSnitch as endpoint.

• Define the Data Center and Rack that this node run on. The default settings:

    dc=DC1
    rack=RAC1
  

• Here, we are using two machines so change the file based on our setup.

    dc=DC1
    rack=RAC1
    dc=DC2
    rack=RAC1
  

• Add the above datacenter and rack in the /etc/cassandra/cassandra-topology.properties

• It look like this

    Data Center One
    175.56.12.105=DC1:RAC1
    175.50.13.200=DC1:RAC1
    175.54.35.197=DC1:RAC1
  
    120.53.24.101=DC1:RAC2
    120.55.16.200=DC1:RAC2
    120.57.102.103=DC1:RAC2
  

• Change the datacenter and rack into your ipaddress in that file.

    ipaddress1=DC1:RAC1
    ipaddress2=DC2:RAC1
  

• Use this command to check the data center name, and rack is set into your machine.

    nodetool status. ![](/blog-assets/content/images/2016/07/cassandra.png)
  

• Its show the status of the node.

    `UN` - define your node in up status.
    `DN` - define your node in down status.
  

Step 3- Configure cassandra.yaml in dc2

• Its assumed that you have installed cassandra in the two nodes (dc1, dc2).

• Once you installed cassandra in your machine, then you need to change the cassandra yaml file in (dc2 as well).

Repeat the process in as many racks/datacenters.

Conclusion

These are the very simple steps to setup cassandra replications in several nodes.

To Deploy your app

[Image unavailable in the original archive]