Kafka Streams：“更新的集群元数据版本”运行了很长时间

Question

在过去的几个小时里，我的 Kafka Streams 工作试图启动但被Updated cluster metadata version一个非常大的主题所占据：

2018-01-03T13:30:56.871 DEBUG [org.apache.kafka.clients.consumer.KafkaConsumer] - Starting the Kafka consumer
2018-01-03T13:30:56.882 DEBUG [org.apache.kafka.clients.Metadata] - Updated cluster metadata version 1 to Cluster(id = null, nodes = [node2:9092 (id: -2 rack: null), 
  node3:9092 (id: -3 rack: null), node1:9092 (id: -1 rack: null)], partitions = [])
...
2018-01-04T10:10:27.945 DEBUG [org.apache.kafka.clients.NetworkClient] - Sending metadata request (type=MetadataRequest, topics=<ALL>) to node node3:9092 (id: 3 rack: null)
2018-01-04T10:10:27.952 DEBUG [org.apache.kafka.clients.Metadata] - Updated cluster metadata version 704157 to Cluster(id = DmwzpJYRTEGrm8MfjcpXLw, nodes = [node3:9092 (id: 3 rack: null), 
  node1:9092 (id: 1 rack: null), node2:9092 (id: 2 rack: null)], partitions = [Partition(topic = raw-messages, partition = 9, leader = 3, replicas = [3], isr = [3]), 
  Partition(topic = raw-messages, partition = 8, leader = none, replicas = [2], isr = []), 
  Partition(topic = raw-messages, partition = 7, leader = 1, replicas = [1], isr = [1]), 
  Partition(topic = raw-messages, partition = 6, leader = 3, replicas = [3], isr = [3]), 
  Partition(topic = raw-messages, partition = 5, leader = none, replicas = [2], isr = []), 
  Partition(topic = raw-messages, partition = 4, leader = 1, replicas = [1], isr = [1]), 
  Partition(topic = raw-messages, partition = 3, leader = 3, replicas = [3], isr = [3]), 
  Partition(topic = raw-messages, partition = 2, leader = none, replicas = [2], isr = []), 
  Partition(topic = raw-messages, partition = 1, leader = 1, replicas = [1], isr = [1]), 
  Partition(topic = raw-messages, partition = 0, leader = 3, replicas = [3], isr = [3])])

这是一个带有新应用程序 ID 的新 Kafka Streams 作业，它应该从头开始（偏移 = 0）。

有几个事件可能导致这种行为：

• 我们从 Confluent Platform 3.2.0 迁移到 Confluent Platform 4.0.0
• 数据中心停电导致Kafka集群严重崩溃

我的问题是，我应该让它运行吗？目前，工作在metadata version 704157，我可以找出或估计版本总数吗？