OK so I misread your comment.  All it really means is that someone closed a 
connection to the thrift DRPC server.  You should look in the logs of the 
workers to see if any of them have been killed and relaunched recently.


- Bobby


On Wednesday, June 28, 2017, 10:03:22 AM CDT, sam mohel <[email protected]> 
wrote:

I haven't see any exception except I wrote here . Connection reset by peer and 
this in droc log file . And in local mode just stopped without any message 

On Wednesday, June 28, 2017, Bobby Evans <[email protected]> wrote:
> It should have been in the exception.  You have not included enough 
> information from the logs to be able to actually debug this.  I am just 
> guessing from similar exceptions I have seen in the past.
>
>
> - Bobby
>
>
> On Wednesday, June 28, 2017, 9:23:22 AM CDT, sam mohel <[email protected]> 
> wrote:
>
> @Bobby Thanks for replying , please how can i check if ip-address and port if 
> it has another worker ? i'm on ubuntu 14.04 LTS 
> On Wed, Jun 28, 2017 at 4:15 PM, Bobby Evans <[email protected]> wrote:
>
> Connection reset by peer typically means that another worker was rescheduled 
> some place else and that worker closed it's connection to this host.  If it 
> did come from the worker then you should see the IP address + port of the 
> other worker and see if it was rescheduled.  If this was because something 
> else closed the connection then it is hard to tell what is happening.
> With DRPC it is not guaranteed to be processed.  If the message is not 
> processed in a timely manor you do need to retry it.
>
>
> - Bobby
>
>
> On Wednesday, June 28, 2017, 9:00:08 AM CDT, sam mohel <[email protected]> 
> wrote:
>
> Thanks for replying and help . i tried to increase  worker.heap.memory. mb: 
> 2048 but not working . DRPC stopped working . I wonder why it is stopped and 
> data set i used smaller than first one !! my data set are tweets and i'm 
> working on processing it . l tried local mode  but  also not working result 
> size stopped in size 57.7 KB. Is there any thing i should share it ? 
> On Wed, Jun 28, 2017 at 2:56 PM, Navin Ipe <navin.ipe@searchlighthealth. com> 
> wrote:
>
> @Sam: You've provided very less information for us to help you. Prima facie, 
> if you have allocated very less memory for your topologies, Storm is 
> obviously running out of memory, the spouts and bolts are restarting which 
> causes the Connection reset by peer error.
> The solution is to allow Storm to use more RAM (assuming there is more RAM).
>
> int RAM_IN_MB = 2048;
> Use stormConfig.put(Config. TOPOLOGY_WORKER_MAX_HEAP_SIZE_ MB, RAM_IN_MB);
>
> If you provide more details of the error, when it happens and what your 
> program is trying to accomplish, the others on this forum would be able to 
> help you better.
>
>
> On Wed, Jun 28, 2017 at 3:30 PM, sam mohel <[email protected]> wrote:
>
> Is there any help.please ?
>
> On Wednesday, June 28, 2017, sam mohel <[email protected]> wrote:
>> I submitted two topologies in production mode . First one has a data set 
>> with size 215 MB and worked well  and gave me the results . Second topology 
>> has a data set with size 170 MB with same configurations but stopped worked 
>> after some times and didn't complete its result 
>> The error i got is drpc log file 
>>     TNonblockingServer [WARN] Got an IOException in internalRead!
>>     java.io.IOException: Connection reset by peer
>> I couldn't figure where is the problem as it supposed to work well as second 
>> data set is smaller in size 
>
>
> --
> Regards,
> Navin
>
>

Reply via email to