[BUG] Join Failure - Mixed Version cluster #19272
Description
Describe the bug
During node-join, leader sends the latest cluster-state to joining nodes for validation. The cluster manager compresses and caches this serialized cluster-state (JoinHelper#serializedState) to reuse across multiple joining nodes.
Version-specific attributes introduced in 2.17 and 3.1 make the cached serializedState incompatible across different opensearch versions. This can cause join failures in mixed-version clusters (e.g., 2.19 and 3.1) when nodes fail to deserialize incompatible cluster state.
Related component
No response
To Reproduce
- Create a cluster with mixed version (2.19 and 3.1)
- Bounce the nodes alternatively between 2.19 and 3.1 in quick succession
- Verify the logs for error
Expected behavior
We need to introduce additional logic to ensure that the node joining the cluster can deserialize the cached state before sending it as part of validation request. If it is not compatible, the version specific cluster-state needs to be regenerated based on joining node version.
Additional Details
Plugins
Please list all plugins currently enabled.
Screenshots
If applicable, add screenshots to help explain your problem.
Host/Environment (please complete the following information):
- OS: [e.g. iOS]
- Version [e.g. 22]
Additional context
Add any other context about the problem here.