f

kafka常见错误整理

一、 flume正常运行，但是kafka接受不了消息，flume日志打印如下

2019-01-18 10:13:44,762 (SinkRunner-PollingRunner-DefaultSinkProcessor) [INFO - kafka.utils.Logging$class.info(Logging.scala:68)] Fetching metadata from broker id:0,host:hadoop01,port:9092 with correlation id 9 for 1 topic(s) Set(weblog)
2019-01-18 10:13:44,763 (SinkRunner-PollingRunner-DefaultSinkProcessor) [INFO - kafka.utils.Logging$class.info(Logging.scala:68)] Connected to hadoop01:9092 for producing
2019-01-18 10:13:44,765 (SinkRunner-PollingRunner-DefaultSinkProcessor) [INFO - kafka.utils.Logging$class.info(Logging.scala:68)] Disconnecting from hadoop01:9092
2019-01-18 10:13:44,766 (SinkRunner-PollingRunner-DefaultSinkProcessor) [INFO - kafka.utils.Logging$class.info(Logging.scala:68)] Disconnecting from hadoop01:9092
2019-01-18 10:13:44,769 (SinkRunner-PollingRunner-DefaultSinkProcessor) [INFO - kafka.utils.Logging$class.info(Logging.scala:68)] Connected to hadoop01:9092 for producing
2019-01-18 10:13:44,771 (SinkRunner-PollingRunner-DefaultSinkProcessor) [WARN - kafka.utils.Logging$class.warn(Logging.scala:83)] Produce request with correlation id 11 failed due to [weblog,0]: kafka.common.NotLeaderForPartitionException

根据NotLeaderForPartitionException可以知道是broker中没有leader导致的异常。我是先创了一个weblog的topic，然后宕掉了其中一台，刚好那台主机是leader，所以导致异常。解决办法：删掉weblog节点，去除宕掉的那台。或者重新启动宕掉的那台。