1.44.6

• [Java] Update RecoverPlan after standby snapshot replication completes with the replicated snapshot entries.

1.46.7

• [Java] Prevent an untethered subscription re-joining the stream at an old position.

1.44.5

• [Java] Prevent an untethered subscription re-joining the stream at an old position.

1.45.1

• [Java] Prevent an untethered subscription re-joining the stream at an old position.

1.46.6

• [Java] Fix a performance regression in ControlTransportPoller/DataTransportPoller when the number of transports is larger than five, i.e. the bytesReceived count was zero and as a result IdleStrategy.idle() was invoked on each duty cycle.
• [Java] Assert that synchronous ControlSession calls are only allowed from a conductor thread.
• [C] Check subscribed sessions before removing a stream interest before removing by stream_id.
• [Java/C] Do not fail fast on I/O exceptions in the send/receive path, i.e. process remaining transports/publications/destinations in the same duty cycle.
• [Java] Make it clear that appointed leader config is a testing feature only.
• [Java] Keep publishing position updates from canvass and into nominate state so other cluster node members can take action in extended election timeout duration.
• [Java] Track total snapshot duration on the follower nodes.
• [Java] Upgrade to Agrona 1.23.1.

1.44.4

• [C] Check subscribed sessions before removing a stream interest before removing by stream_id.
• [Java] Add end-to-end test for session interest fix.

1.46.5

• [Java] Delete segment files without renaming so that the Archive conductor thread will not be blocked for a long period of time.
• [Java] Remove expensive file existence checks when scheduling segment file deletion.

1.46.4

• [C] Align loss and SMs handling between Java and C media drivers.

1.44.3

• [C] Align loss and SMs handling between Java and C media drivers.

1.46.2

• [Java] Fix ClusterBackup issue whereby upon restart it could end up reading data from the previously started replay thus failing to recover the log.