'properties.transaction.timeout.ms' = '3' 配置的太短了，30s
transactionalId 就过期了。 估计 都来不去启动吧
官网的原文
https://ci.apache.org/projects/flink/flink-docs-release-1.13/docs/connectors/datastream/kafka/
Semantic.EXACTLY_ONCE mode relies on the ability to commit transactions
that were started before
您好：kafka在exactly-once语义下,从svaepoint恢复报错。初步排查认为是kafka事务使用了旧的epoch。请问这个问题怎么处理？
//todo 通过配置传进来
env.setParallelism(1);
env.enableCheckpointing(60L, CheckpointingMode.EXACTLY_ONCE);
// checkpoint的清除策略(即使任务被显示地取消也会保留checkpoint)
env.getCheckpointConfig()
