Say in Mongo we have one primary and two replica node(simple replica no sharding). Read and write request will be handled only primary node and replica will just backup the data in case of failover.

My understanding/question about replication(not sharding) in Cassandra

Can we have only one node(say Node_1) serving read and write requests while other two simply working as replica nodes(these nodes just takes the replica request from Node_1 , no real time request from user ) like in Mongo considering replication factor as 3 ?

Per my understanding this should not be possible in cassandra, as in cassandra , replica nodes are nodes are equal peers , so it is mandatory that any of three nodes has to serve write and read request.

Is my understanding correct ?


Yes, your understanding is correct - this allows evenly distribute load between nodes & to handle more traffic.

Why would you need replicas if they just idle? If you're afraid of inconsistent data, then just set correct consistency levels to get strong consistency...

  • You said Why would you need replicas if they just idle? Like in MongoDB, replica will just keep backup of data in case of failover. So there also they are kind of idle too . Is n't it ? So i thought is it possible in cassandra also where replica node just work as backup(and not take anr read,write request from user) in case of failover. – user3198603 Jun 21 '18 at 1:28
  • In Cassandra this is not required - because of its architecture all nodes are equal. In Mongo the need for leader is because of its architecture... – Alex Ott Jun 21 '18 at 9:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.