How to Reset Kafka Consumer Group Offset

how_to_reset_kafka_consumer_group_offset.md

Kafka 0.11.0.0 (Confluent 3.3.0) added support to manipulate offsets for a consumer group via cli kafka-consumer-groups command.

1. List the topics to which the group is subscribed

kafka-consumer-groups --bootstrap-server <kafkahost:port> --group <group_id> --describe

Note the values under "CURRENT-OFFSET" and "LOG-END-OFFSET". "CURRENT-OFFSET" is the offset where this consumer group is currently at in each of the partitions.

2. Reset the consumer offset for a topic (preview)

kafka-consumer-groups --bootstrap-server <kafkahost:port> --group <group_id> --topic <topic_name> --reset-offsets --to-earliest

This will print the expected result of the reset, but not actually run it.

3. Reset the consumer offset for a topic (execute)

kafka-consumer-groups --bootstrap-server <kafkahost:port> --group <group_id> --topic <topic_name> --reset-offsets --to-earliest --execute

This will execute the reset and reset the consumer group offset for the specified topic back to 0.

4. Repeat 1 to check if the reset is successful

Note

• The consumer group must have no running instance when performing the reset. Otherwise the reset will be rejected.
• There are many other resetting options, run kafka-consumer-groups for details
  • --shift-by <positive_or_negative_integer>
  • --to-current
  • --to-latest
  • --to-offset <offset_integer>
  • --to-datetime <datetime_string>
  • --by-duration <duration_string>
• The command also provides an option to reset offsets for all topics the consumer group subscribes to: --all-topics

For ones who want to do the same for Kafka 0.10 you may download Kafka 1.0 binaries which include the script with this new offset reset capabilities, it will work for Kafka 0.10 as well:

wget http://apache-mirror.rbc.ru/pub/apache/kafka/1.0.0/kafka_2.11-1.0.0.tgz
tar -xvf kafka_2.11-1.0.0.tgz
cd kafka_2.11-1.0.0/bin

# Check what offsets is flume-archive currently on:
./kafka-consumer-groups.sh --bootstrap-server <kafka:port> --group <group> --describe

# Reset the offset
./kafka-consumer-groups.sh --bootstrap-server <kafka:port>--group <group>--topic <topic> --reset-offsets --to-earliest --execute

# Check what offsets is flume-archive currently on again to make sure that offset is now resetted properly:
./kafka-consumer-groups.sh --bootstrap-server <kafka:port> --group <group> --describe

# Current offset for all topic partitions should be now reduced. I.e. LOG-END-OFFSET should be bigger than CURRENT-OFFSET

thx for instruction :) Do you know why it is so hard to achieve in most libraries? Did you do any research about it? I have checked that in kafka-go, sarama ( both golang) and spring-kafka - there is no easy way to reset offset while using consumer groups. I just have problem with understanding if its dangerous to reset offset or there are some other technical issues.

More description:
https://stackoverflow.com/questions/49360325/kafka-consumer-group-loses-not-commited-messages

thanks daniilyar. but we're facing below error:

Error: Executing consumer group command failed due to The broker only supports OffsetFetchRequest v1, but we need v2 or newer to request all topic partitions.

Seems like version incompatibility here to execute commands with 1.0 binaries.

Anyone know if there's a way to do this on Kafka 0.9.0?

Apparently my latests and my earliests offsets have the same values, and are equal to LOG-END-OFFSET. I can not manage to go back to the beginning of the messages.

kafka .10 does not support --rest-offsets but this can be achieved using java apis.
Please have a look.

https://gist.github.com/mduhan/0e0a4b08694f50d8a646d2adf02542fc

Can anyone tell me how to do this for kafka 1.0.0

The API is all the same since 0.11. This includes 1.0.0, 2.0, and above until anyone mentions otherwise.

Applying 1.0 to 0.10 didn't work for me either. Same as @Joginder22:
Error: Executing consumer group command failed due to The broker only supports OffsetFetchRequest v1, but we need v2 or newer to request all topic partitions.

I would like to make a small adjustment about how to reset a topic partition

./kafka-consumer-groups --bootstrap-server <kafkahost:port> --group <group_id> --topic <topic_name>[:<partition number>] --reset-offsets --to-earliest --execute

kafka1.1 doesn't seem to support reset by --to-datetime. it can effect --to-earliest but do not effect --to-datetime

Anyone knows if it's possible to reset the group offset using librdkafkacpp APIs? Thanks

Greetings all,

Just leaving a note here to save everyone using a Kafka cluster protected by SSL time. In case you are getting the following OOM error when running kafka-consumer-groups:

[2020-02-06 09:31:40,554] ERROR Uncaught exception in thread 'kafka-admin-client-thread | adminclient-1': (org.apache.kafka.common.utils.KafkaThread)
java.lang.OutOfMemoryError: Java heap space
	at java.nio.HeapByteBuffer.<init>(HeapByteBuffer.java:57)
	at java.nio.ByteBuffer.allocate(ByteBuffer.java:335)
	at org.apache.kafka.common.memory.MemoryPool$1.tryAllocate(MemoryPool.java:30)
	at org.apache.kafka.common.network.NetworkReceive.readFrom(NetworkReceive.java:112)
	at org.apache.kafka.common.network.KafkaChannel.receive(KafkaChannel.java:436)
	at org.apache.kafka.common.network.KafkaChannel.read(KafkaChannel.java:397)
	at org.apache.kafka.common.network.Selector.attemptRead(Selector.java:653)
	at org.apache.kafka.common.network.Selector.pollSelectionKeys(Selector.java:574)
	at org.apache.kafka.common.network.Selector.poll(Selector.java:485)
	at org.apache.kafka.clients.NetworkClient.poll(NetworkClient.java:539)
	at org.apache.kafka.clients.admin.KafkaAdminClient$AdminClientRunnable.run(KafkaAdminClient.java:1152)
	at java.lang.Thread.run(Thread.java:748)

Counterintuitively, you are not running out of memory, but rather this arises as a result of not providing SSL credentials. You need to create a .properties file and pass it into each call to the binary via the --command-config option. In our case, the template for this properties file is:

ssl.endpoint.identification.algorithm=https
sasl.mechanism=PLAIN
request.timeout.ms=20000
bootstrap.servers=BOOTSTRAP
retry.backoff.ms=500
sasl.jaas.config=org.apache.kafka.common.security.plain.PlainLoginModule required username="KEY" password="PASSWORD";
security.protocol=SASL_SSL

Hope this saves someone time in the future.

Thanks @muscovitebob!

Thanks @muscovitebob, saved my day.

I find it rediculous that there is no --partition option or similar, I MUST fix an offset for only one partition without modifying the others.

@scottcarey it's possible, check my message

https://gist.github.com/marwei/cd40657c481f94ebe273ecc16601674b#gistcomment-3082013

--topic <topic_name>[:<partition number>]

Is it possible to reset offsets on an active consumer groups?

Is it possible to reset offsets on an active consumer groups?

@ghenadiibatalski Nope, you will be asked to stop consumers

All, anyone faces timeout, Timed out waiting to send the call

All, anyone faces timeout, Timed out waiting to send the call

Resolved: by using --timeout 10000 param option.

Does anyone know the minimum permissions required to reset offsets? Every doc about kafka ACL operations that I've looked through is vague about permissions for resetting offsets.

For Kafka 2.4.0 (Commit:77a89fcf8d7fa018) after kafka docker restart, reset offsets to earliest is refusing to reset to 0.

docker exec -it kafka /kafka/bin/kafka-consumer-groups.sh \
                                --bootstrap-server kafka:9092 \                 
                                --group <mygroup> \                     
                                --reset-offsets \
                                --to-earliest \
                                --all-topics \
                                --execute
                                
GROUP                          TOPIC                          PARTITION  NEW-OFFSET
<mygroup>                     topic1                                  0        1649

Thanks for this! I come here about once every couple weeks because I can never remember these commands

@dcguim @dkurzaj - I'm seeing the same behaviour. Offset are stuck to their initial value. Have you managed to solve this ?

I have not managed to solve this problem, nor found the specific configuration for this. What I can do is delete and recreate the problematic topic.

I actually do not remember, sorry...

Apparently my latests and my earliests offsets have the same values, and are equal to LOG-END-OFFSET. I can not manage to go back to the beginning of the messages.

@dkurzaj that is because your topic doesn't contain messages, and the offset can only be reset to the first message present in the topic at max