hei, one question regarding KafkaSource and .setStartingOffsets method. What would happen if I use this ? .setStartingOffsets(OffsetsInitializer.earliest()). Does this mean that every new redeployment of the job will read from the beginning of the topic? or the offset will be saved in the state and will pick up next time that we redeploy the job from a savepoint?

It depends if you're enabling checkpointing or not

I have checkpoint enabled

So when you would restart from your checkpoint, Flink would resume where it left of and ignore the setting

same applies to savepoint I guess right? whether you resume from checkpoint or savepoint right?

Correct, the snapshotting mechanism is the same

makes me wonder what the difference is then in between *.setStartingOffsets(OffsetsInitializer.earliest()). and .setStartingOffsets(OffsetsInitializer.co*mmitedOffsets(OfffsetStrategy.EARLIEST))

I believe in the second one, you could also start from a previously committed offset, and only if those are not available you would start from earliest

right, thanks

Get an OffsetsInitializer which initializes the offsets to the committed offsets. Use the given OffsetResetStrategy to initialize the offsets if the committed offsets does not exist.

I was right 😅