You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, running a CascadingDataSink with a DOP > 1 yields in empty results because the job commit logic removes the result files for some reason. We had a similar issue in Flink's Hadoop compatibility wrappers for Hadoop OutputFormats and solve it by moving the commiting to the JM via FinalizeOnMaster (see FLINK-1139).
Right now, the DOP of data sinks is fixed to 1 in FlinkFlowStep.translateSink().
The text was updated successfully, but these errors were encountered:
Currently, running a CascadingDataSink with a DOP > 1 yields in empty results because the job commit logic removes the result files for some reason. We had a similar issue in Flink's Hadoop compatibility wrappers for Hadoop OutputFormats and solve it by moving the commiting to the JM via
FinalizeOnMaster
(see FLINK-1139).Right now, the DOP of data sinks is fixed to 1 in
FlinkFlowStep.translateSink()
.The text was updated successfully, but these errors were encountered: