[CASSANDRA-19847] Create a fuzz test that randomizes topology changes, cluster actions, and CQL operations

• Type: Improvement
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 5.1
• Component/s: Test/fuzz
• Change Category: Quality Assurance
• Complexity: Normal
• Platform: All

Source Control Link

https://github.com/apache/cassandra/commit/fdabc1f9774b1c06d68a0acbf8a470e45c787eec

Test and Documentation Plan

new test

• Assignee: David Capwell
• Reporter: David Capwell

Authors

David Capwell

Reviewers

Alex Petrov

• Created: 21/Aug/24 17:03
• Updated: 04/Sep/24 18:08
• Resolved: 04/Sep/24 17:54

Description

This is a spinoff from CASSANDRA-19769 to lower the patch size

To help validate the cluster state it is good to put a cluster into random situations to make sure it still works correctly. CASSANDRA-19769 Added such a test and found several issues with TCM (all fixed) and can be used to help find more issues as we expand the scope of what it can touch.

I propose we have a test that randomizes changes to the topology, performs cluster operations such as repair, and performs CQL operations (for this patch using Harry)

Issue Links

• is blocked by: CASSANDRA-19872 Handle losing CMS status while committing a transformation
• relates to: CASSANDRA-19768 nodetool assassinate of a CMS voting member should be allowed and should remove it from the CMS group
• relates to: CASSANDRA-19782 Host replacements no longer fully populate system.peers table
• relates to: CASSANDRA-19761 When JVM dtest is shutting down, if a new epoch is being committed the node is unable to shut down
• relates to: CASSANDRA-19769 CEP-15: (Accord) sequence EpochReady.coordinating to allow syncComplete to be learned from newer epochs
• relates to: CASSANDRA-19705 Reconfigure CMS after move/bootstrap/replacement
• links to: GH: accord trunk
• links to: GH: cep-15-accord
• links to: GH: trunk
• links to: GitHub Pull Request #119
• links to: GitHub Pull Request #3509