owner info in zk is not correct

Hi, We are using kafka 0.7. 2 brokers, each broker has 10 partitions for one topic 3 consumers in one consumer group, each consumer create 10 streams. Today, when we want to rollout new service. After we restart one consumer we find exceptions and warning. kafka.common.ConsumerRebalanceFailedE

Re: owner info in zk is not correct

Hello Yonghui, In 0.7 the consumer rebalance logic is distributed and in some corner cases such as soft-failure-caused-consecutive rebalances some consumer may consider the rebalance as complete while others are still trying the rebalance process. You can check the GC logs on your consumer to veri

Kafka 0.8's ConsumerConnector.close() hangs if ZK is unavailable and autocommit is enabled

Hi all, I'm using Kafka 0.8 and I've ran into an issue with ConsumerConnector. Steps to reproduce: 1. Start single-broker Kafka cluster with auto.create.topic.enable set to "true" 2. Start ConsumerConnector on topic (which does not yet exist) with auto.offset.reset set to "smallest". 3. Produce s

Loss of Leader in Kafka

Hi, I am hitting a strange exception while creating a topic in Kafka - Steps to generate this- 1. Created a topic multipartition_test with 2 partitions and 2 replicas 2. Added some data to this topics and verified data is coming up for both partitions 3. Deleted the topic. Checked only the zookeepe

Re: question about mirror maker

if placing mirror maker in the same datacenter as target cluster, it/consumer will talks to zookeeper in remote/source datacenter. would it more susceptible to network problems? As for the problem commit offset without actually producing/writing msgs to target cluster, it can be solved by disablin

IndexOutOfBoundsException error

Hi all, I am seeing some corrupt data on my 0.7.2 Kafka cluster. Every once in awhile I'll get the following message: Exception in thread "kafka-consumer" java.lang.IndexOutOfBoundsException at java.nio.Buffer.checkIndex(Unknown Source) at java.nio.HeapByteBuffer.get(Unknown Sour

Re: question about isr

You're probably right that it has to be GC given the CMS activity I saw in the log but I didn't see a single concurrent mode failure, which bothers me that we had this happen anyway... also the ZK timeout is set to a very large number... I dunno, seems weird. I will see what I can do about gett

Re: question about isr

Could you try 0.8.1.1, which fixed some bugs related to controller hanging during shutdown. For ZK session expiration, the #1 cause is GC. We have also seen that transient issues with the underlying storage (e.g. raid controller reset) can also cause ZK session expiration. As for the data loss in

Re: question about isr

Another question - If I start broker 1 back up now, with his replicas now out of sync, will he automatically catch up with the leaders or do I have to do something to catch him up? --Ian On May 12, 2014, at 10:49 AM, Jun Rao wrote: > Could you try 0.8.1.1, which fixed some bugs related to co

Re: question about isr

0.8.1.1 fixed an issue when the controller could hang during soft failure. Out-of sync replicas will catch up from the leader automatically, once restarted. Thanks, Jun On Mon, May 12, 2014 at 12:37 PM, Ian Friedman wrote: > You're probably right that it has to be GC given the CMS activity I

Kafka producer in CSharp

Hi I have kafka broker running (kafka_2.9.1-0.8.1.1) All is working. One project requires producer is written in CSharp I am not dot net programmer but I managed to write simple producer code using https://github.com/kafka-dev/kafka/blob/master/clients/csharp/README.md the code ... using Sys

Re: IndexOutOfBoundsException error

Could you run the DumpLogSegments tool to see if there are indeed corrupted messages in the broker log? Thanks, Jun On Mon, May 12, 2014 at 4:08 PM, Xuyen On wrote: > Hi all, > > I am seeing some corrupt data on my 0.7.2 Kafka cluster. Every once in > awhile I'll get the following message: >

Re: kafka broker failed to recovery from ZK failure

The controller log in broker 1 is too late. Could you send its log around 2014-05-12 21:24:37? Thanks, Jun On Mon, May 12, 2014 at 5:02 PM, Steven Wu wrote: > This is a three-node cluster. broker 0 lost connection to ZK. broker 1 > does seem to take the controller role. but broker 0 stuck in

Re: Loss of Leader in Kafka

Delete topic doesn't quite work in 0.8.1. We recently fixed it in trunk. Could you give it a try and see if you see the same issue? Thanks, Jun On Mon, May 12, 2014 at 9:39 AM, Kashyap Mhaisekar wrote: > Hi, > I am hitting a strange exception while creating a topic in Kafka - > Steps to genera

Re: question about mirror maker

As far as Zookeeper goes, any time you have network communication you have the change of a problem. I would rather have the network issue on the consumer side, rather than the producer side. I would certainly prefer to have the offsets committed only after the message is produced (based on the ack

CSharp librari and Producer Closing socket for because of error (kafka.network.Processor),java.nio.BufferUnderflowException

Hi I have kafka broker running (kafka_2.9.1-0.8.1.1) All is working. One project requires producer is written in CSharp I am not dot net programmer but I managed to write simple producer code using https://github.com/kafka-dev/kafka/blob/master/clients/csharp/README.md the code ... using Sys