[fix][broker] Avoid false WARN log when replicator is already terminated

[Denovo1998]

Fixes #xyz

Main Issue: #xyz

PIP: #xyz

Motivation

AbstractReplicator.closeProducerAsync can race with termination. In the post-close callback, the state check mixed two different state snapshots, so a normal transition to Terminated could be misclassified and logged as an unexpected WARN. This creates noisy logs and makes concurrent shutdown behavior harder to reason about.

Modifications

• Fixed the terminated-state branch condition in AbstractReplicator to consistently use setDisconnectedRes when evaluating Terminating/Terminated.
• Kept the close/termination state machine behavior unchanged, and only corrected the condition used for log-path selection.
• Added testCloseProducerAsyncNoWarnWhenAlreadyTerminated to cover the race where another thread sets the replicator state to Terminated while producer.closeAsync() is still in progress.
• Asserted that no unexpected WARN log is emitted and that the expected INFO log path is used.

Verifying this change

• Make sure that the change passes the CI checks.

Documentation

• doc
• doc-required
• doc-not-needed
• doc-complete

Matching PR in forked repository

PR in forked repository: apache#25259