Scenario:

• Start up a topic where some partitions are on broker 0, some are on broker 1
• Move the partitions to a single broker (0)
• Kill broker 1, as it is no longer relevant to the topics currently in service

You will get this sort of log (from the example consumer in this case):

• 1428602624.709 RDKAFKA-3-FAIL: rdkafka#consumer-0: 10.42.0.144:9092/1: Failed to connect to broker at 10.42.0.144:9092: Operation timed out
• 1428602624.709 RDKAFKA-3-ERROR: rdkafka#consumer-0: 10.42.0.144:9092/1: Failed to connect to broker at 10.42.0.144:9092: Operation timed out
• 1428602625.715 RDKAFKA-3-FAIL: rdkafka#consumer-0: 10.42.0.144:9092/1: Failed to connect to broker at 10.42.0.144:9092: Host is down
• 1428602625.715 RDKAFKA-3-ERROR: rdkafka#consumer-0: 10.42.0.144:9092/1: Failed to connect to broker at 10.42.0.144:9092: Host is down
• 1428602718.698 RDKAFKA-3-FAIL: rdkafka#consumer-0: 10.42.0.144:9092/1: Failed to connect to broker at 10.42.0.144:9092: Operation timed out
• 1428602718.698 RDKAFKA-3-ERROR: rdkafka#consumer-0: 10.42.0.144:9092/1: Failed to connect to broker at 10.42.0.144:9092: Operation timed out

Producers also do this.