Cassandra 3 node configuration

0
votes

I am new to Cassandra, use latest Cassandra 3.10. I have 3 nodes to link to participate in Cassandra. Cluster name Test Cluster same as three nodes. Same Datacenter dc1 ,Rack as rack1 and snitch as GossipingPropertyFileSnitch used .It Configures

Node A:

-seeds : "A,B,C address"

listen_address & rpc_address are same to A node ip address

Node B:

-seeds : "A,B,C address"

listen_address & rpc_address are same to B node ip address

Node C:

-seeds : "A,B,C address"

listen_address & rpc_address are same to C node ip address

What i am do possibility here listed

i) suppose if A node is failure get data from node B and C .

ii) If any one or two node failure get data from another node. How to configure these nodes.

I have use Simple Strategy with replication factor 3 has used. If node failure get node from another node data retrieve so, seeds address or mistaken? Briefly explain what to do.

1
cassandra
You could use DataStax OpsCenter to configure and install Cassandra for you, or read the DataStax documentation – OneCricketeer
Yes i have read but little confused in seed address and replication factor – P.Surendar
Replication factor = how many copies of data you need. In a small cluster, you don't need a factor of 3... Seed address only needs to be a single (external) ip address in a small cluster. All nodes must have the same seeds (at least that's how I configure it) – OneCricketeer
Ok. I tried three nodes data replicated. If sample java thread run on single node A replicate B and C node. So all node have data. Seed point to all other node is right? my configuration is right. – P.Surendar
Think of the seeds as the "master" or the "brain". Each config points to the same list of servers. If you can start the services and run nodetool status, and see servers, then it works – OneCricketeer

1 Answers

0
votes

Answering your questions:

If a Node A goes down, then you want to fetch data from node B and C.
If one or two node goes down, you want to fetch data from other node.

To achieve the above, the replication factor which you have configured is enough to handle the node failure. The wrong configuration is having all your nodes be seed node.

A seed node is used to bootstrap other nodes, So usually first node is started first in a data center as a seed node. Suppose you have 2 data centers. Then you should have 2 seed nodes, as mentioned in below datastax docs: http://docs.datastax.com/en/cassandra/3.0/cassandra/initialize/initSingleDS.html

As per your last comment, you have mentioned "schema version mismatch detected". Which is means all your nodes are not in same cluster. Check the schema using nodetool when all your nodes are running

nodetool describecluster

This should give nodes schema version. All nodes should be same schema version.
So if any one node does not have same version then restart the node till the schema version is same.

Once you fix this schema error, you will be able to create keyspace.