spark從kafka獲取數據很慢

環境：spark 1.4 + kafka 0.9  3臺相同配置主機 每臺主機一個broker、standalone方式運行spark集羣，每臺機器2個workerhtml TOPIC分佈 Topic: MessageBody Partition: 0 Leader: 0 Replicas: 0 Isr: 0 Topic: MessageBody Partition: 1 Leader: 1 Re

>>阅读原文<<