Cassandra Admin

No. Apache Cassandra is an open-source project and the ASF does not issue administrator certifications. Cassandra Admin credentials come from independent training providers and vendors. Check the specific certifying body's site for its syllabus, its current fees and scheduling, and whether the exam maps to a particular Cassandra release.

Multiple-choice and scenario-based multiple-choice, delivered online or at a test centre depending on the provider. Scenarios typically present a cluster symptom and ask for the best next action or the most likely root cause. Distractors are usually plausible-but-wrong operational moves, so read for the qualifier words - 'first', 'most likely', 'safest'.

Straightforward if your experience is broad; hard if it is narrow. People who only add nodes and run repairs tend to fail on JVM and compaction questions; people who only tune JVM tend to fail on repair, security and topology. Audit your weak domains against the published weights before you book, and lab the ones you avoid at work.

Start with architecture and the distributed ring, because partitioning, replication and consistency questions become easy once token ownership is intuitive. Then data modelling and performance tuning, which lean on the same mental model. Leave repair and security for last - they are more self-contained and reward memorising concrete parameters and defaults.

Rarely. Documentation gives you definitions but not the diagnostic reflexes the scenarios test. You need to have watched nodetool tpstats drop blocked tasks, seen what a compaction backlog does to latency, and restored a snapshot. Reading plus a three-node lab is the realistic minimum.

Confusing replication factor with consistency level; assuming QUORUM on a two-DC cluster means what it means on one; recommending nodetool repair when the real answer is fixing a tombstone or compaction problem; and forgetting that ALL availability depends on every replica being up. Also: answering the textbook fix instead of the safest first step.

Providers generally allow retakes, sometimes after a waiting period and usually at a reduced or full fee. The exact retake policy, waiting window and cost are set by the certifying organization, so read its candidate handbook rather than relying on forum reports. Use the score report's domain breakdown to target the retake.

Validity periods differ by provider - some issue lifetime credentials, others require renewal every few years as Cassandra majors advance. Check the certifying body's site for the current validity term and renewal route, which is often continuing education or sitting the current exam version rather than a separate recertification test.

Usually it targets a recent stable major release, and the syllabus states which. Version matters: compaction, streaming and authentication defaults have changed across majors. Confirm the mapped version on the provider's site, and make sure your lab cluster runs that version before you rely on memory for defaults.

It depends on your baseline. A working DBA with a lab habit can usually close the gaps in a few focused weeks of evenings; someone new to Cassandra should plan considerably longer, because the ring and read/write path take repetition to internalise. Track progress by domain, not by hours logged.