In this case I have just published the message via a producer and
connecting to the same broker list to consume.

On Thu, Dec 10, 2015, 2:18 PM Jason Gustafson <ja...@confluent.io> wrote:

> Hi Kevin,
>
> At the moment, the timeout parameter in poll() really only applies when the
> consumer has an active partition assignment. In particular, it will block
> indefinitely to get that assignment. If there are no brokers to connect to
> or if you accidentally point it to an 0.8 cluster, it will probably block
> forever. This is most likely what is happening in your test case. This is a
> known issue, but surprisingly difficult to handle in a reasonable way due
> to the fact that the consumer is single-threaded. We are working on this
> problem, but for now, the best solution is to use wakeup() to break from
> the poll() from an external thread.
>
> -Jason
>
> On Thu, Dec 10, 2015 at 11:50 AM, Kevin Carr <kscar...@gmail.com> wrote:
>
> > I am configuring the way it says in the api doc.
> >
> > Topic is 1 partition
> > I set the group.id, and subscribe to the topic.
> >
> > If I call poll(100) the function never returns.  I have to stop the unit
> > test, because it never stops.  I have waited for up to 10 minutes.
> >
> > Any Ideas?
> >
> > Thanks for your help.
> >
>

Reply via email to