cqlsh

Start CQL shell

$ cqlsh

$ cqlsh [hostname] [9042]

$ cqlsh -u [username] -p [password]

$ cqlsh -f [script.cql]

$ cqlsh -e "SELECT * FROM [keyspace].[table]"

$ nodetool status

$ nodetool info

$ nodetool repair

cqlsh [options] [host] [port]
nodetool [options] command [args]

cqlsh is the Cassandra Query Language shell for executing CQL statements interactively or from scripts. nodetool manages Cassandra node operations.
CQL resembles SQL with some differences. Keyspaces are like databases; tables have partition keys and clustering columns. Queries must include partition key for efficient lookups.
nodetool status shows cluster health with Up/Down/Normal/Leaving/Joining states and token ownership. nodetool repair ensures data consistency across replicas.
Snapshots create hard-linked backups of SSTables for point-in-time recovery. cleanup removes data after topology changes. compact forces SSTable compaction.
Node lifecycle operations: drain prepares for shutdown, decommission removes node from cluster streaming data to others.

~/.cassandra/cqlshrc

User-specific cqlsh settings including connection defaults, display formatting, and authentication credentials.

Main Cassandra configuration file controlling cluster behavior, data directories, network settings, and replication.

-u, --username user

Authentication username.

Authentication password.

Initial keyspace.

Execute CQL file.

Execute single CQL statement.

Use SSL connection.

Connection timeout.

Query timeout.

status

Show cluster node status.

Node information.

Token ring information.

Run anti-entropy repair.

Remove data not belonging to node.

Force compaction.

Flush memtables to SSTables.

Take snapshot backup.

Decommission the node.

Drain node for shutdown.

Cluster information.

CQL is not SQL; avoid joins and complex queries. Wide partitions impact performance. Repair is resource-intensive; schedule appropriately. Decommission before removing nodes. Authentication default is off.

Apache Cassandra was initially developed at Facebook for inbox search and open-sourced in 2008. It became an Apache project in 2009 and top-level in 2010. Cassandra combines Dynamo's distribution model with Bigtable's data model. DataStax was founded in 2010 to provide commercial support. Version 4.0 (2021) brought major improvements after extensive testing.

scylladb(1), cockroach(1), mongo(1)