kafka-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From wenxing zheng <wenxing.zh...@gmail.com>
Subject NotEnoughReplicasException on __transaction_state after bring down one of kakfa broker
Date Wed, 17 Apr 2019 04:26:59 GMT
Dear all,

*The problem*
Our kafka cluster is running with 7 brokers and 3 days ago, we brought down
node 8 for upgrading the HW. After that, we found the error log below
saying: org.apache.kafka.common.errors.NotEnoughReplicasException: Number
of insync replicas for partition.

Before shutting down, our application worked well with enabling
transaction. But after the operation, the application stopped working.

Please kindly advice what would be the problem here.


April 14th 2019, 17:03:52.421
[2019-04-14 17:03:45,045] TRACE [Controller id=1 epoch=42] Sending
UpdateMetadata request PartitionState(controllerEpoch=42, leader=6,
leaderEpoch=138, isr=[6], zkVersion=259, replicas=[8, 6, 7, 1, 2],
offlineReplicas=[8]) to brokers Set(1, 5, 2, 6, 7, 4) for partition
__transaction_state-25 (state.change.logger)
April 14th 2019, 17:03:52.533
[2019-04-14 17:03:51,574] ERROR [ReplicaManager broker=6] Error processing
append operation on partition __transaction_state-25
(kafka.server.ReplicaManager)
org.apache.kafka.common.errors.NotEnoughReplicasException: Number of insync
replicas for partition __transaction_state-25 is [1], below required
minimum [3]
*Server version*
Kafka version: 2.0.1
Client version: 2.0.1
JDK: 1.8

Thank you very much,
Regards,

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message