Closed nathan-szz closed 1 year ago
Here is my configuration information: starrocks: 'sink.max-retries' = '10', 'sink.buffer-flush.interval-ms' = '30000', 'sink.properties.strip_outer_array' = 'true', 'sink.properties.format' = 'json', 'sink.properties.ignore_json_size' = 'true', 'sink.properties.connectTimeout' = '10000', 'sink.connect.timeout-ms' = '10000', 'sink.version'='V2'
flink:
execution.checkpointing.interval: 30000
state.checkpoints.num-retained: 20
execution.checkpointing.mode: EXACTLY_ONCE
How did you measure the delay? Did the flink job have back pressure?
What's the version of connector?
What's the version of connector?
flink-connector-starrocks-1.2.7_flink-1.15
I found reason what cause this bug. it’s not the connector .
I discovered that in a multi-table synchronization task, such as multiple 'insert into ...' statements, when testing data insertion into Starrocks, there is a delay of 5 minutes for one particular table. However, when performing a synchronization task for a single table alone, the delay returns to normal, at around 1 minute.