Hi,
I am on Digi Ocean FYI and not dealing with the issue of private and public
IP address. I killed all kafka servers and rebuilt but have the same zk
cluster. One would think on kafka 0.8.2 that if I rebuild kafka, register
with ZK all would be right with the world but guess not. I have new broker
id's 1 and 2 where prior it was a random integer.
Here is my producer.properties file which is the same on both servers
metadata.broker.list=104.xxx.xxx.xxx:9092,198.xxx.xxx.xxx:9092
producer.type=sync
compression.codec=none
serializer.class=kafka.serializer.DefaultEncoder
So from the below...it looks to me that I have a clear leader.
bin/kafka-topics.sh --describe --zookeeper xxx.xxx.xxx.xxx:2181 --topic
topic-impression-production
Topic:topic-impression-production PartitionCount:2
ReplicationFactor:2 Configs:
Topic: topic-impression-production Partition: 0 Leader: 21729
Replicas: 25906,21729 Isr: 21729
Topic: topic-impression-production Partition: 1 Leader: 21729
Replicas: 21729,25906 Isr: 25906,21729
In zookeeper this is what I see in exhibitor:
brokers
ids
1
2
topics
topic-impression-production
partitions
0
state
1
state
When i try and write I get the below. What am i missing?
ubuntu@ubuntu:/var/kafka$ bin/kafka-console-producer.sh --broker-list
1xxx.xxx.xxx.xxx:9092 --topic topic-impression-production
te[2016-01-08 18:38:19,530] WARN Property topic is not valid
(kafka.utils.VerifiableProperties)
stfadfad
[2016-01-08 18:38:28,652] WARN Error while fetching metadata partition
0 leader: none replicas: isr: isUnderReplicated: false for
topic partition [topic-impression-production,0]: [class
kafka.common.LeaderNotAvailableException]
(kafka.producer.BrokerPartitionInfo)
[2016-01-08 18:38:28,653] WARN Error while fetching metadata partition
1 leader: none replicas: isr: isUnderReplicated: false for
topic partition [topic-impression-production,1]: [class
kafka.common.LeaderNotAvailableException]
(kafka.producer.BrokerPartitionInfo)
