Hi Wangsan,
what you are observing is the intended behaviour of Flink in the case of
cancelling a job. The assumption is that if a job is cancelled, then there
should be no clean shut down. This entails that the Tasks are interrupted
in order to terminate them as fast as possible.
Thus, the
Hi all,
We are currently using BucketingSink to save data into HDFS in parquet format.
But when the flink job was canceled, we always got Exception in BucketingSink’s
close method. The detailed exception info is as below:
[ERROR] [2017-09-26 20:51:58,893]
