Apache Cassandra 5.0 on Rocky Linux 9 by cloudimg. Wide column NoSQL, PasswordAuthenticator enabled, per-VM superuser password rotation at first boot, cloudimg keyspace pre-created. Apache 2.0 licensed, no per-node fees.
## Apache Cassandra 5.0 on Rocky Linux 9 by cloudimg
Apache Cassandra 5.0, the latest major release of the wide-column NoSQL database that powers Netflix, Apple, Instagram, Discord, and thousands of production deployments. Built on Rocky Linux 9 for Microsoft Azure and maintained by cloudimg.
Why Choose cloudimg?
* 24/7 Expert Support with guaranteed 24 hour response. Contact support@cloudimg.co.uk
* Production Ready from Launch Pre-configured, security-patched, validated
* Azure Native Integration Azure Linux Agent, cloud-init, Gen2 Hyper-V, TrustedLaunch
* Per VM Superuser Rotation A fresh 32-char random password is generated for the cassandra superuser at first boot and written to /stage/scripts/cassandra-credentials.log (mode 0600 root only). No two VMs share credentials.
What is Included
* Apache Cassandra 5.0 from the official redhat.cassandra.apache.org/50x DNF repo
* OpenJDK 17 JRE headless (Java 17 required)
* PasswordAuthenticator + CassandraAuthorizer + CassandraRoleManager enabled by default
* cassandra-firstboot.service generating per-VM cassandra superuser + cloudimg user passwords, creating the cloudimg keyspace, writing /stage/scripts/cassandra-credentials.log
* Python 3.9 + cassandra-driver venv at /opt/cqlsh-venv
* CQL native protocol on TCP 9042 (firewalld opens 9042 externally; 7000/7001/7199 stay private)
* Data dir /var/lib/cassandra (NOT /mnt — Azure ephemeral resource disk does not survive SIG capture)
* Rocky Linux 9 base with latest security patches
* 24/7 cloudimg support with guaranteed 24 hour response SLA
Cassandra 5.0 Capabilities
* Wide-column data model with tunable consistency (ONE / QUORUM / ALL / LOCAL_QUORUM)
* Linear horizontal scaling — add nodes to grow capacity, no downtime
* Multi-DC replication via NetworkTopologyStrategy and GossipingPropertyFileSnitch
* SAI (Storage-Attached Index) for secondary index queries — production-ready in 5.0
* Vector data type (5.0 GA) for ANN search workloads — pairs naturally with embedding-based RAG pipelines
* Materialized Views, UDFs, UDAs, change-data-capture (CDC)
* Apache 2.0 licensed; no per-node, per-CPU, or per-GB fees
Recommended Virtual Machine Sizes
* Standard_B2s (2 vCPU, 4 GB RAM) — single-node dev/test
* Standard_D4s_v5 (4 vCPU, 16 GB RAM) — small production cluster nodes
* Standard_E16s_v5 (16 vCPU, 128 GB RAM) — large production with attached Premium SSDs