If your topology has saved Kafka offset in your zookeeper it will start 
processing from that otherwise It checks spoutConfig.forceFromStart set to true 
in this case it will try to fetch data from the beginning of the queue i.e 
kafka.api.EarliestTime() . If none of the above matches it will pick users 
spoutConfig.startOffsetTime.

"kafka.api.OffsetRequest.EarliestTime() finds the beginning of the data in the 
logs and starts streaming from there, kafka.api.OffsetRequest.LatestTime() will 
only stream new messages.”

If you stopping and re-deploying the topology make sure you used the same name 
as KafkaSpout uses topology name to store and retrieve the offsets from 
zookeeper.

-- 
Harsha



On March 9, 2015 at 7:30:38 AM, Tousif ([email protected]) wrote:

If your topology has saved Kafka offset in your zookeeper it will start 
processing from that otherwise It checks spoutConfig.forceFromStart set to true 
in this case it will try to fetch data from the beginning of the queue i.e 
kafka.api.EarliestTime() . If none of the above matches it will pick users 
spoutConfig.startOffsetTime.

"kafka.api.OffsetRequest.EarliestTime() finds the beginning of the data in the 
logs and starts streaming from there, kafka.api.OffsetRequest.LatestTime() will 
only stream new messages.”

If you stopping and re-deploying the topology make sure you used the same name 
as KafkaSpout uses topology name to store and retrieve the offsets from 
zookeeper.

-- 
Harsha

Reply via email to