Cassandra cluster know how

The problem:

Cassandra was throwing write and read timeouts like

Cassandra timeout during read query at consistency LocalQuorum (1 replica(s) responded over 2 required)

The issue was mismatch in schema versions. In order to check the schema version following command have to be executed: nodetool describecluster http://i.imgur.com/EEvVs4H.png

If any of the nodes is unreacheable you have to execute same command from that cassandra.

The solution:

Restart nodes one by one until all of them have same schema version as described in http://docs.datastax.com/en/landing_page/doc/landing_page/cstarTroubleshooting/schemaDisagree.html. Execute nodetool repair just in case (might be good idea to flush first nodetool flush)

Setup:

Cassandra nodes: 3
Replication factoer: 3
Read consistency level: LocalQuorum
Write consistency level: LocalQuorum
Replication strategy: network_topology

and

Cassandra nodes: 3
Replication factoer: 1
Read consistency level: LocalOne
Write consistency level: LocalOne
Replication strategy: simple