You probably also want to check the request time jmx (in particular
FetchRequest time) on the broker to see if the broker is healthy.
Thanks,
Jun
On Fri, Sep 26, 2014 at 3:02 PM, Jagbir Hooda wrote:
> Hi Jun,
>
> Thanks for the info, we'll keep you posted on the above topic.
> Surprisingly in
Hi Jun,
Thanks for the info, we'll keep you posted on the above topic.
Surprisingly in broker controller logs we also see somewhat similar
exceptions at the same line BoundedByteBufferReceive.scala:54. (Not
sure if there is any correlation here.)
8<--
[2014-09-
You can enable some trace/debug level logging to see if the thread is
indeed hanging in BoundedByteBufferReceive.
Thanks,
Jun
On Wed, Sep 24, 2014 at 8:30 AM, Jagbir Hooda wrote:
> Hi Jun,
>
> Thanks for looking into it. We use the following steps to start the
> consumer.
>
> 1) Create consume
Hi Jun,
Thanks for looking into it. We use the following steps to start the consumer.
1) Create consumer connector
ConsumerConnector consumerConnector =
kafka.consumer.Consumer.createJavaConsumerConnector(createConsumerConfig())
2) Create message streams
List> streamList =
consumerConnecto
We allocate a new BoundedByteBufferReceive for every fetch request. Are you
using SimpleConsumer directly? It seems it's started by the high level
consumer through the FetchFetcher thread.
Thanks,
Jun
On Mon, Sep 22, 2014 at 11:41 AM, Jagbir Hooda wrote:
> Note: Re-posting the older message f
Note: Re-posting the older message from another account due to
formatting issues.
Folks,
Recently in one of our SimpleConsumer based client applications (0.8.1.1),
we spotted a very busy CPU with almost no traffic in/out from the client
and Kafka broker (1broker+1zookeeper) (the stack trace is
I'm sorry about the formatting issues below:-(I need to stop using hotmail as
the hotmail is mangling the message formatting:-(I'll try re-posting from my
gmail address.
Jagbir
> From: jsho...@hotmail.com
> To: users@kafka.apache.org
> Subject: Busy CPU while negotiating co
Folks,
Recently in one of our SimpleConsumer based client applications (0.8.1.1),we
spotted a very busy CPU with almost no traffic in/out from the clientand Kafka
broker (1broker+1zookeeper) (the stack trace is attached at the end).
The busy thread was invoked in a while loop anchored at the read
