https://ops.docs.72602.space/demo/plugins/flink-s3-f3-multiple/index.html Normally, Flink only can access only one S3 endpoint during the runtime. But we need to process some files from multiple minio simultaneously. So I modified the original flink-s3-fs-hadoop and enable flink to do so. StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment(); env.enableCheckpointing(5000L, CheckpointingMode.EXACTLY_ONCE); env.setParallelism(1); env.setStateBackend(new HashMapStateBackend()); env.getCheckpointConfig().setCheckpointStorage("file:///./checkpoints"); final FileSource<String> source = FileSource.forRecordStreamFormat( new TextLineInputFormat(), new Path( "s3u://admin:ZrwpsezF1Lt85dxl@10.11.33.132:9000/user-data/home/conti/2024-02-08--10")) .build(); final FileSource<String> source2 = FileSource.forRecordStreamFormat( new TextLineInputFormat(), new Path( "s3u://minioadmin:minioadmin@10.101.16.72:9000/user-data/home/conti")) .build(); env.fromSource(source, WatermarkStrategy.noWatermarks(), "file-source") .union(env.fromSource(source2, WatermarkStrategy.noWatermarks(), "file-source2")) .print("union-result"); env.execute(); original usage example using default flink-s3-fs-hadoop, the configuration value will set into Hadoop configuration map. Only one value functioning at the same, there is no way for user to operate different in single one job context. Hugo en