Thanks for the good suggestion. I will log in the fail method and try. Since 
this problem happens intermittently, its difficult to reproduce.

> On Sep 27, 2015, at 8:21 PM, Madan <[email protected]> wrote:
> 
> Yes, you are right. This should be enabled by default. Can you check from the 
> logs that the particular message is getting replayed. 
> Another sinpler way is to  print the message at spout's fail method to see if 
> its getting invoked by the timeout.
> Regards,
> Madan
> 
>> On Sep 28, 2015 08:33, "Sudha Subramanian" <[email protected]> wrote:
>> Hi Madan,
>> 
>> Are you referring to topology.enable.message.timeouts.
>> 
>> I have not configured explicitly. I'm guessing the default value should be 
>> true? Is that not the case?
>> 
>> Thanks,
>> Sudha
>> 
>> 
>>> On Sat, Sep 26, 2015 at 10:47 AM, Madan <[email protected]> wrote:
>>> Hi,
>>> Have you enable message timeout. The Spout should retry emitting the 
>>> message, if its not acked by the bolt.
>>> Reards,
>>> Madan
>>> 
>>>> On Sep 26, 2015 19:44, "Sudha Subramanian" <[email protected]> wrote:
>>>> Hi,
>>>> 
>>>> I have a spout that emits a message on a stream. Since its on a 
>>>> development environment, I have a single instance of bolt to receive the 
>>>> message from stream and process it. The bolt is expected to ack the 
>>>> message to spout.  
>>>> 
>>>> It works fine majority of the times except when the bolt sometimes does 
>>>> not receive the emitted message from the spout. I have to kill and 
>>>> redeploy the topology to fix this.
>>>> 
>>>> Can someone share some insight on what might cause this behavior? 
>>>> 
>>>> Thanks,
>>>> Sudha

Reply via email to