GridGain 9.0.1 Release Notes

In this release, client metrics for Java and .NET clients were expanded with support for continuous query metrics. You can use these metrics to track the number of received continuous query events received and sent.

For Java clients, make sure to enable metrics in configuration:

You can then use the standard JMX monitoring tools to track metrics.

For .NET clients, the metrics are exposed automatically and can be tracked without client modifications.

DDL requests, such as CREATE TABLE, take a few seconds each to complete. Because of that, large database initialization scripts may take longer than expected. This will be addressed in an upcoming release.

Some scenarios may see lower performance when moving to GridGain 9. This is a temporary limitation, and the performance in these scenarios will be improved in upcoming versions.

When a partition loses majority of its copies, it becomes read-only. This behavior is required for split-brain protection. Because of this, in a distribution zone with 2 data replicas, losing one node may lead to partial unavailability.

To achieve both split-brain protection and full availability with one node down, use 3 or more replicas.

Upcoming versions will add a high-availability mode that support full availability with one node down when using 2 replicas.

Complex Data Center Replication topologies (for example, the ones involving cycles) of 3 or more data centers are not supported. This will be addressed in an upcoming release.

The following features of GridGain 8 are not available in this version, and will be added in upcoming versions: