hazelcast/hazelcast-python-client

[TRACKING ISSUE] [API-1531] Send client state to cluster in re-connections

hz-devops-test opened this issue · 0 comments

The tracking issue for the Java side PR.

See hazelcast/hazelcast#22164 for details.


In split-brain scenarios, the cluster id stays the same for different halves of the split.

When the client disconnects from the first half it was connected to and reconnects to the other half, from the point of view of the client, it connects to the same cluster.

However, it might very well be the case that, the client has sent some state to the first half of the cluster, and that state must be replicated to the other half when the client reconnects to it.

This is especially needed for Compact schemas and user code deployment classes.

In this PR, we are sending the client state to the cluster after reconnections, regardless it is connected back to possibly the same cluster with the same id or not.