Hi Dan, Usually broadcast state needs more network buffers, the network buffer used to exchange data records between tasks would request a portion of direct memory[1], I think it is possible to get the “Direct buffer memory” OOM errors in this scenarios. Maybe you can try to increase taskmanager.memory.framework.off-heap.size <https://nightlies.apache.org/flink/flink-docs-master/docs/deployment/config/#taskmanager-memory-framework-off-heap-size> and taskmanager.memory.task.off-heap.size <https://nightlies.apache.org/flink/flink-docs-master/docs/deployment/config/#taskmanager-memory-task-off-heap-size> .
[1] https://nightlies.apache.org/flink/flink-docs-master/docs/deployment/memory/mem_setup_tm/#detailed-memory-model Best, Yanfei Dan Hill <quietgol...@gmail.com> 于2022年10月21日周五 15:39写道： > Hi. My team recently added broadcast state to our Flink jobs. We've > started hitting this OOM "Direct buffer memory" error. Is this a common > problem with broadcast state? Or is it likely a different problem? > Thanks! - Dan >
