Hello Flink friends !
I have two questions on
checkpointing:
• In the documentation
here, 3 strategies are mentioned for the checkpointing strategy: at most once, at least once, exactly once. But in the checkpointing parameters documentation
here, for the parameter "execution.checkpointing.mode", only EXACTLY_ONCE and AT_LEAST_ONCE. Is at most once available ?
• Tolerable checkpoint failure number ("execution.checkpointing.tolerable-failed-checkpoints"). If that is set to 1, does that mean that one checkpoint can be skipped (i.e. it fails and the job still moves on to the next checkpoint interval) ? Or does it mean that a failed checkpoint will be retried once ?
Thanks for your help