Your partitions are balanced, but is your data being evenly written across all
the partitions? How are you producing data? Are you producing them with keys?
Is it possible that the majority of the messages being written to just a few
partitions, and so the brokers for those partitions are seeing
Hello Tao,
For your case maybe you can monitor the following jmx as well (see
http://kafka.apache.org/documentation.html#monitoring):
kafka.server:type=BrokerTopicMetrics,name=BytesOutPerSec
When a broker cannot properly respond to requests it will be much smaller
compared with other brokers.
G
Hello,
1) When the fetch offset is out of range, broker will send back the
corresponding error code and the new consumer should reset its offset
automatically according to the reset policy (config "auto.offset.reset"),
by default it is set to "latest", i.e. then log end offset. If you do not
have
Hello,
Did you have compression turned on and batching (in terms of #.messages)?
In that case the whole compressed message set is treated as a single
message on the broker and hence could possibly exceed the limit.
In newer versions we have changed the batching criterion from #.messages to
bytes,
Thanks Stevo,
Feel free to paste your findings in KAFKA-2937, we can re-open that ticket
if necessary.
Guozhang
On Fri, Mar 4, 2016 at 4:38 AM, Stevo Slavić wrote:
> Hell Apache Kafka community,
>
> I'm still investigating an incident; from initial findings topic deletion
> doesn't seem to wor
Hello Alexis,
Could you share your findings about the command line tool? We can try to
resolve if there's any issues.
Guozhang
On Fri, Mar 4, 2016 at 3:13 PM, Alexis Midon <
alexis.mi...@airbnb.com.invalid> wrote:
> The command line tool that ships with Kafka is error prone.
>
> Our standard pr
Hi Mansi,
1) subscribe is a lazy call: only the subsequent poll() will do the leg
work of talking to broker, registering the subscription, getting assigned
partitions and issuing fetch requests to the partition leaders.
2) currently the timeout value for poll() does not count for the time
elapsed
Guozhang
Thanks for the explanation !!
Mansi.
Sent from my iPhone
> On Mar 5, 2016, at 3:38 PM, Guozhang Wang wrote:
>
> Guozhang
