Zhijiang created FLINK-16537: -------------------------------- Summary: Implement ResultPartition state recovery for unaligned checkpoint Key: FLINK-16537 URL: https://issues.apache.org/jira/browse/FLINK-16537 Project: Flink Issue Type: Sub-task Components: Runtime / Checkpointing, Runtime / Network Reporter: Zhijiang Fix For: 1.11.0
During recovery process for unaligned checkpoint, the partition state should also be recovered besides with existing operator states. The ResultPartition would request buffer from local pool and then interact with ChannelStateReader to fill in the state data. The filled buffer would be inserted into respective ResultSubpartition queue in normal way. It should guarantee that op can not process any inputs before finishing all the output recovery to avoid mis-order issue. Refer to more details by [https://docs.google.com/document/d/16_MOQymzxrKvUHXh6QFr2AAXIKt_2vPUf8vzKy4H_tU/edit] -- This message was sent by Atlassian Jira (v8.3.4#803005)