Replication in cassandra using – PropertyFileSnitch (NetworkTopologyStrategy)

Replication is the process of storing copies of data on multiple nodes to ensure reliability and fault tolerance. Replication Strategy 1. SimpleStrategy -- simple single data center clusters. 2. NetworkTopologyStrategy --cluster deployed across multiple data centers.

 Here I am working on 5 node Cassandra cluster. CC1 192.168.56.138 (seed 1) CC2 192.168.56.142 CC3 192.168.56.144 CC4 192.168.56.143 (seed 2) CC5 192.168.56.145 Step 1: Edit the cassandra.yaml file in all nodes.(endpoint_snitch: SimpleSnitch -- it’s default )

1.1 endpoint_snitch: SimpleSnitch  replace with endpoint_snitch: PropertyFileSnitch.
1.2 Each DataCenter we have one seed node , that we have to update in cassandra.yaml file
seed_provider:
    -seed: “ 198.168.56.138,198.168.56.143”
Node : Do it on all nodes
Step 2: Edit this file as /conf/cassandra-topology.properties in all nodes
Note: Remove or comment the line’s b/w (# Cassandra Node IP=Data Center:Rack  ----- # default for unknown nodes)
NOTE:   Replace as you configuration  
NOTE:  3 node in Data center 1 and 2 node in Data center 2
# DATA_CENTER :1
192.168.56.138=DC1:RAC1
192.168.56.142=DC1:RAC1
192.168.56.144=DC1:RAC1

# DATA_CENTER :2
192.168.56.143=DC2:RAC1
192.168.56.145=DC2:RAC1

# default for unknown nodes
default=DC1:RAC1
Step 3: Start Cassandra in all node’s(./cassandra –f)
Check with (./nodetool status) .. it’s show as you specified on about properties file

STEP :4 Test scenario : 1 Test the setting … How the replication factor working in cluster with 2 datacenters.
4.1 create Keyspace with NetworkTopologyStrategy … here I am maintain data in Datacenter 1 all nodes(3), Datacenter 2 all nodes(2).

cqlsh> create keyspace webapp_test with replication= { 'class'    :'NetworkTopologyStrategy','DC1':3, 'DC2':2};

4.2 . This Command Show the key space structure.

cqlsh> desc keyspace webapp_test ;
CREATE KEYSPACE webapp_test  WITH replication = {
  'class': 'NetworkTopologyStrategy',
  'DC2': '2',
  'DC1': '3'
};

4.3 Create Column Family Account.

cqlsh> use webapp_test ;
cqlsh:webapp_test > create table account (accname text PRIMARY KEY , url text, username text);
cqlsh:webapp_test > desc tables;

account

4.4 Insert data

cqlsh> INSERT INTO webapp_test .account(accname,username,url) values('1','TLlive','www.tllive.com');
cqlsh> INSERT INTO webapp_test .account(accname,username,url) values('2','TLlive_2','www.tllive2.com');
cqlsh> INSERT INTO webapp_test .account(accname,username,url) values('3','ESPN','www.espn.com');

cqlsh> select * from webapp_test .account;

 accname | url             | username
---------+-----------------+----------
       3 |    www.espn.com |     ESPN
       2 | www.tllive2.com | TLlive_2
       1 |  www.tllive.com |   TLlive

(3 rows)

cqlsh> quit;

4.4 It will show the data distributed nodes in the cluster using( getendpoints)

It will show the data distributed nodes in the cluster using( getendpoints)
user@cc1:~/cassandra/bin$ ./nodetool getendpoints webapp_test  account accname
192.168.56.144
192.168.56.145
192.168.56.143
192.168.56.138
192.168.56.142 
Note: Data replicated in Dc1 3 node & DC2 2 nodes)
138,142,144---dc1
143,145 –dc2
----------------

Test scenario : 2
4.2.1 : Alter the keyspace webapp_test as

cqlsh> Alter keyspace webapp_test  with replication= { 'class' : 'NetworkTopologyStrategy','DC1':2, 'DC2':1};
cqlsh> desc keyspace webapp_test  ;

CREATE KEYSPACE webapp_test  WITH replication = {
  'class': 'NetworkTopologyStrategy',
  'DC2': '1',
  'DC1': '2'
};

cqlsh> quit;
user@cc1:~/cassandra/bin$ ./nodetool getendpoints webapp_test  account accname
192.168.56.144
192.168.56.145
192.168.56.138
Note: Data replicated in Dc1 1 node & DC2  2 nodes)
138---dc1
143,145 –dc2

Test scenario : 3

cqlsh> desc keyspace web_app1;

CREATE KEYSPACE web_app1 WITH replication = {
  'class': 'SimpleStrategy',
  'replication_factor': '2'
};
cqlsh> Alter  KEYSPACE web_app1 WITH replication = {'class' :'NetworkTopologyStrategy','DC1':2, 'DC2':2};
cqlsh> desc keyspace web_app1;

CREATE KEYSPACE web_app1 WITH replication = {

  'class': 'NetworkTopologyStrategy',
  'DC2': '2',
  'DC1': '2'

user@cc1:~/cassandra/bin$ ./nodetool repair
It will run in all nodes sequential fashion 
user@cc1:~/cassandra/bin$ ./nodetool getendpoints web_app1 users username
192.168.56.144
192.168.56.143
192.168.56.138
192.168.56.145

Acknowledgements

Thank you Rakesh Budagam and David Letchumanan for this post.