Page History

...

Once a particular Startpoint is applied to the starting offset of a system-stream-partition in a task instance, it is subsequently removed at the next offset commit. 

Failure Scenarios

SamzaContainer is restarted before the first offset commit

• The Startpoint will be applied to the starting offset again upon restart. However, this is identical to the same failure scenario for Checkpoints.

JobCoordinator is restarted before all Startpoints are applied

• Similar to the above failure scenario, except across multiple containers since restarting a JobCoordinator restarts all associated SamzaContainers.

Startpoint Models

WIP

...