loserwang1024 commented on code in PR #27177: URL: https://github.com/apache/flink/pull/27177#discussion_r2488402219
########## docs/content/docs/dev/datastream/sources.md: ########## @@ -54,6 +54,16 @@ The Data Source API supports both unbounded streaming sources and bounded batch The difference between both cases is minimal: In the bounded/batch case, the enumerator generates a fixed set of splits, and each split is necessarily finite. In the unbounded streaming case, one of the two is not true (splits are not finite, or the enumerator keeps generating new splits). + +**Split Reassignment On Recovery** + +Under normal circumstances, once the *SplitEnumerator* assigns *Splits* to *SourceReaders*, these *splits* are not reassigned to other readers again. When the source is recovering from a failure, the *splits* from the saved state will be added back to the readers immediately. + +When a source implements the `SupportsSplitReassignmentOnRecovery` interface, the recovery process behaves differently. Review Comment: The expectation on the source is : "On Recovery, instead of immediately reassigning the *splits* back to the same *SourceReaders*, all *splits* are collected and added back to the *SplitEnumerator*." When startup(not failover), there is no split to recover. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
