Discussion:
Kafka error - append operation on partition
Faisal Durrani
2018-11-16 02:47:04 UTC
Permalink
Hi Users,

Apologies if this question has been asked before but i have not been able
to find the answer.

I am seeing this error in the Kafka logs :

[ReplicaManager broker=1001] Error processing append operation on
partition TABLE-0 (kafka.server.ReplicaManager)
org.apache.kafka.common.errors.UnknownProducerIdException: Found no record
of producerId=39002 on the broker. It is possible that the last message
with the producerId=39002 has been removed due to hitting the retention
limit.

This is the 2nd time we are seeing this error , we changed the retention
limit to a larger value when we got the error the first time but its back
now. We have 5 node Kafka cluster with Oracle golden gate as producer. The
topic TABLE as 5 partition but at this point the data is not getting
distributed on the cluster because( because Oracle is sending the same
kafka.key for all the messages) . The compression type is set to 'producer'
and at GG its set to gzip.

Can you let me know what can we do to avoid getting this error as it is
causing the GG to go Abend.

Regards,
Faisal
Faisal Durrani
2018-11-16 06:37:24 UTC
Permalink
I went through the logs segment of the topic and i see that the data from
the producerID still exists in the topic and the new messages received are
well with in the retention period ( 336 hr).

Starting offset: 359
offset: 359 position: 0 CreateTime: 1541667838856 isvalid: true keysize: 39
valuesize: 500 magic: 2 compresscodec: GZIP producerId: 39002
producerEpoch: 0 sequence: 0 isTransactional: false headerKeys: [] key:
${currentDate[yyyy-mm-dd hh:MM:ss.SSS]} payload: {"table":
offset: 360 position: 405 CreateTime: 1542334682507 isvalid: true keysize:
39 valuesize: 501 magic: 2 compresscodec: GZIP producerId: 41003
producerEpoch: 0 sequence: 0 isTransactional: false headerKeys: [] key:
${currentDate[yyyy-mm-dd hh:MM:ss.SSS]} payload: {"table":
offset: 361 position: 405 CreateTime: 1542334682507 isvalid: true keysize:
39 valuesize: 500 magic: 2 compresscodec: GZIP producerId: 41003
producerEpoch: 0 sequence: 1 isTransactional: false headerKeys: [] key:
${currentDate[yyyy-mm-dd hh:MM:ss.SSS]} payload: {"table":"

I would be grateful if someone can help me find the root cause
Post by Faisal Durrani
Hi Users,
Apologies if this question has been asked before but i have not been able
to find the answer.
[ReplicaManager broker=1001] Error processing append operation on
partition TABLE-0 (kafka.server.ReplicaManager)
org.apache.kafka.common.errors.UnknownProducerIdException: Found no record
of producerId=39002 on the broker. It is possible that the last message
with the producerId=39002 has been removed due to hitting the retention
limit.
This is the 2nd time we are seeing this error , we changed the retention
limit to a larger value when we got the error the first time but its back
now. We have 5 node Kafka cluster with Oracle golden gate as producer. The
topic TABLE as 5 partition but at this point the data is not getting
distributed on the cluster because( because Oracle is sending the same
kafka.key for all the messages) . The compression type is set to 'producer'
and at GG its set to gzip.
Can you let me know what can we do to avoid getting this error as it is
causing the GG to go Abend.
Regards,
Faisal
Loading...