Hi
I posted this to IRC but maybe someone here has seen this before
Hello
I'm having some weird issues with kafka 0.9 and can't find anythign in jira
I'm running Kafka inside Docker/Kubernets
All works fine when I deploy But after a while I get the following in the
publisher
WARN: Error while fe
Hello
following doc @
https://cwiki.apache.org/confluence/display/KAFKA/System+Tools#SystemTools-SimpleConsumerShell
i tried to print messages using the command
./kafka-run-class.sh kafka.tools.SimpleConsumerShell —-max-messages 1
--no-wait-at-logend —-print-offsets --partition 17 --offset 7644
Hi All,
We are using Kafka Producer in our web application. Things are fine when it
is up, but in case it goes down, the log is over populated with the
following error :
2016-03-07 12:37:42,813 WARN [kafka-producer-network-thread | producer-1]
[Selector] [line : 276 ] - Error in I/O with /10.4
Hi Jason,
Thanks for your detailed explain.
Face to this situation, I wanna discuss more a little bit,
I try two approach to avoid it in Kafka 0.9.0.1, and both work correctly.
1. Using subscribe(topics, listener) and
implements onPartitionsAssigned(partitions) , it manually run
consumer.commitS
Found the issue. My publisher was not assigning a key to all messages.
On 07/03/2016 14:40, "Rakesh Vidyadharan" wrote:
>Hello,
>
>We are using Kafka 0.8.2.2 and have modified most of our topics to use log
>compaction and a shorter retention.ms equivalent to 24 hours for those topics.
> We
Hi,
We are using Kafka high level consumer , and we are able to successfully
consume messages but the zookeeper connections keep expiring and
reestablishing.
I am wondering why are there no heartbeats to keep the connections alive:
Kafka Consumer Logs
[localhost-startStop-1
Hey Ken,
Whether to use subscribe or assign depends mainly on whether you need to
use consumer groups to distribute the topic load. If you use subscribe(),
then the partitions for the subscribed topics will be divided among all
consumers sharing the same groupId. With assign(), you have to provide
Thanks Guozhang!
No I don't have a way to reproduce this issue. It randomly happens, I am
changing the log level from INFO to trace to see if I can get the exact
message what was sent when this happens.
Could it also be some encoding issue or partial message related?
Thanks,
Fang
On Mon, Mar 7,
I cannot think of an encoding or partial message issue at top of my head
(browsed through 0.8.2.2 tickets, none of them seems related either).
Guozhang
On Tue, Mar 8, 2016 at 11:45 AM, Fang Wong wrote:
> Thanks Guozhang!
>
> No I don't have a way to reproduce this issue. It randomly happens, I
Hi All,
I am new to Kafka, I want to make sure Kafka server is available before I
can do any polling to the messages. What I did is creating a consumer
without subscription and try to access the partitions metadata for a
specific topic. This trick works for the first time. If I take the brokers
do
What we normally do is consumer.poll(0). This connects to the broker,
finds the consumer group, handles partition assignment, gets the
metadata - and then doesn't stick around to actually give you any
data.
Pretty hacky, but we use this all over the place.
Gwen
On Tue, Mar 8, 2016 at 12:59 PM, A
Using the 0.9 consumer, I would like to start consuming at the
beginning or end, without specifying auto.offset.reset.
This does not seem to be possible:
val kafkaParams = Map[String, Object](
"bootstrap.servers" -> conf.getString("kafka.brokers"),
"key.deserializer" -> classOf[St
Hi!
I am quite new to Kafka and pub-sub systems.
I am trying to make a single consumer to consume messages from multiple
topics.
Although I understand that one topic can be sent to multiple consumers via
partitions,
I could not find the part in the documentation that specifies that a single
cons
Hi,
I need your advice how to start server in the next situation:
It fails on startup with FATAL error:
[2016-03-07 16:30:53,495] FATAL Fatal error during KafkaServerStartable
startup. Prepare to shutdown (kafka.server.KafkaServerStartable)
kafka.common.InvalidOffsetException: Attempt to append a
Hi there,
One consumer can indeed consume from multiple topics (and multiple
partitions). For example, see here:
http://kafka.apache.org/090/javadoc/org/apache/kafka/clients/consumer/KafkaConsumer.html#subscribe(java.util.List)
Then, in your poll() loop, you can get the topic and partition from
The characters for the first dash in both --max-messages and
--print-offsets doesn't look like a standard '-', is it possible those
options simply aren't being parsed correctly?
-Ewen
On Tue, Mar 8, 2016 at 12:26 AM, Anishek Agarwal wrote:
> Hello
>
> following doc @
>
> https://cwiki.apache.or
Hi Cody,
The problem with that code is in `seekToBeginning()` followed by
`subscribe(topic)`.
Since `subscribe` call is lazy evaluated, by the time `seekToBeginning()`
is called no partition is assigned yet, and hence it is effectively an
no-op.
Try
consumer.subscribe(topics)
consumer.p
Hi,
Is there a configuration to set poll Interval for a SourceTask using the
Connect API?
The JDBC Connector uses a custom property - poll.interval.ms but is there
an internal property which can be used by different connectors?
Thanks,
Shiti
Thanks ! Stupid textEditor in Mac. I had removed that representation from
"offset" and "partition" forgot the other one.
On Wed, Mar 9, 2016 at 10:17 AM, Ewen Cheslack-Postava
wrote:
> The characters for the first dash in both --max-messages and
> --print-offsets doesn't look like a standard '-
That suggestion doesn't work, for pretty much the same reason - at the
time poll is first called, there is no reset policy and no committed
offset, so NoOffsetForPartitionException is thrown
I feel like the underlying problem isn't so much that seekToEnd needs
special case behavior. It's more tha
20 matches
Mail list logo