Cassandra 5.0: New Features and Ecosystem Evolution

-- Enable vector search extension
-- Requires Cassandra 5.0+

-- Create table with vector column
CREATE TABLE embeddings (
    id UUID PRIMARY KEY,
    document_id UUID,
    embedding VECTOR<FLOAT, 768>,
    created_at TIMESTAMP
);

-- Create vector index (ANN - Approximate Nearest Neighbor)
CREATE CUSTOM INDEX idx_embedding 
ON embeddings USING 'StorageAttachedIndex' 
WITH OPTIONS = {
    'index_name': 'embedding_idx',
    'index_type': 'ANN',
    'similarity_function': 'cosine'
};

-- Search for similar embeddings
SELECT id, document_id, 
       similarity_cosine(embedding, ?)
FROM embeddings
ORDER BY embedding ANN OF ?
LIMIT 10;

-- Enhanced JSON functions
SELECT JSON '{"name": "John", "age": 30}';

-- Parse JSON
SELECT JSON_PARSE('{"name": "John", "age": 30}').name;

-- Convert to JSON
SELECT toJSON(name), toJSON(age) FROM users;

-- Collection functions
SELECT ARRAY_LENGTH(phone_numbers) FROM users;

-- Time functions
SELECT toDate(now());

-- Aggregate improvements
SELECT COUNT(*) FROM users;

-- Improved compaction algorithms
-- Better memory management
-- Reduced CPU overhead

-- Improved network protocol
-- Better handling of large partitions
-- Reduced memory usage

-- Better query planning
-- Improved index usage
-- Reduced read latency

-- New authentication plugins
-- Better password policies

-- Create user with password policy
CREATE ROLE appuser WITH 
    LOGIN = true 
    PASSWORD = 'secure_pass'
    AND PASSWORD EXPIRES IN 90 DAYS;

-- Transparent Data Encryption (TDE)
-- Table-level encryption

-- Enable encryption at rest
ALTER TABLE sensitive_data 
WITH encryption = {
    'key_alias': 'encryption_key'
};

-- Improved cross-DC replication
-- Reduced latency
-- Better conflict resolution

-- Better DC routing
-- Local consistency level options
-- Improved failover handling

# Connect to Astra
from cassandra.cluster import Cluster
from cassandra.auth import PlainTextAuthProvider

cloud_config = {
    'secure_connect_bundle': 'secure-connect-database.zip'
}
auth_provider = PlainTextAuthProvider('client_id', 'client_secret')
cluster = Cluster(cloud=cloud_config, auth_provider=auth_provider)

# k8ssandra.yaml
apiVersion: k8ssandra.io/v1alpha1
kind: K8ssandraCluster
metadata:
  name: my-cluster
spec:
  cassandra:
    serverVersion: "5.0"
    datacenters:
      - metadata:
          name: dc1
        size: 3

# cassandra-datacenter.yaml
apiVersion: cassandra.datastax.com/v1beta1
kind: CassandraDatacenter
metadata:
  name: dc1
spec:
  clusterName: cluster1
  serverType: cassandra
  serverVersion: "5.0"
  size: 3
  storageConfig:
    cassandraDataVolumeClaimSpec:
      storageClassName: standard
      resources:
        requests:
          storage: 10Gi

# 1. Check current version
nodetool version

# 2. Verify cluster health
nodetool status

# 3. Check for running repairs
nodetool compactionstats

# 1. Backup data
nodetool snapshot -t pre_upgrade

# 2. Update Cassandra package
apt-get update
apt-get install cassandra

# 3. Restart nodes (rolling upgrade)
# One node at a time
sudo service cassandra restart

# 4. Verify upgrade
nodetool version

-- Use appropriate partition keys
-- Avoid hot spots
-- Balance partition size

-- Good partition key example
CREATE TABLE events (
    date TEXT,
    event_id TIMEUUID,
    data TEXT,
    PRIMARY KEY ((date), event_id)
);

-- Avoid large partitions
-- Target < 100MB per partition

-- Use prepared statements
-- Batch non-logged updates
-- Monitor compaction