You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@storm.apache.org by Sanjeev Verma <sa...@gmail.com> on 2016/05/23 11:48:52 UTC
Storm topology taking too long to get assigned on worker slot
Hi All
I am observing a strange issue on my storm cluster (6 nodes). After
submitting the topology I can see some 3-7 mins time when worker slot got
occupied.
I am clueless why it is taking so long to occupy the worker slot.I also
observe the same with the work-count topology so it is not topology
specific.
Could somebody provide me some pointers so that I can identify the problem.
Thanks
Re: Storm topology taking too long to get assigned on worker slot
Posted by Abhishek Agarwal <ab...@gmail.com>.
Hi Sanjeev,
can you check the nimbus and supervisor logs after you submit the topology?
Nimbus will set the assignment in zookeeper and supervisor will launch the
worker on its machine if any executor is assigned to that node. Through the
time printed in logs, you could further narrow down the delay.
On Wed, May 25, 2016 at 11:23 AM, Sanjeev Verma <sa...@gmail.com>
wrote:
> looking for some pointers so that can pin point the problem.
>
> Thanks
>
> On Tue, May 24, 2016 at 6:09 PM, Sanjeev Verma <sa...@gmail.com>
> wrote:
>
>> its 10.0.2.4
>>
>> On Tue, May 24, 2016 at 9:45 AM, 马哲超 <ma...@gmail.com> wrote:
>>
>>> Which version do you use?
>>>
>>> 2016-05-23 19:48 GMT+08:00 Sanjeev Verma <sa...@gmail.com>:
>>>
>>>> Hi All
>>>> I am observing a strange issue on my storm cluster (6 nodes). After
>>>> submitting the topology I can see some 3-7 mins time when worker slot got
>>>> occupied.
>>>> I am clueless why it is taking so long to occupy the worker slot.I also
>>>> observe the same with the work-count topology so it is not topology
>>>> specific.
>>>> Could somebody provide me some pointers so that I can identify the
>>>> problem.
>>>>
>>>> Thanks
>>>>
>>>
>>>
>>
>
--
Regards,
Abhishek Agarwal
Re: Storm topology taking too long to get assigned on worker slot
Posted by Abhishek Agarwal <ab...@gmail.com>.
Hi Sanjeev,
can you check the nimbus and supervisor logs after you submit the topology?
Nimbus will set the assignment in zookeeper and supervisor will launch the
worker on its machine if any executor is assigned to that node. Through the
time printed in logs, you could further narrow down the delay.
On Wed, May 25, 2016 at 11:23 AM, Sanjeev Verma <sa...@gmail.com>
wrote:
> looking for some pointers so that can pin point the problem.
>
> Thanks
>
> On Tue, May 24, 2016 at 6:09 PM, Sanjeev Verma <sa...@gmail.com>
> wrote:
>
>> its 10.0.2.4
>>
>> On Tue, May 24, 2016 at 9:45 AM, 马哲超 <ma...@gmail.com> wrote:
>>
>>> Which version do you use?
>>>
>>> 2016-05-23 19:48 GMT+08:00 Sanjeev Verma <sa...@gmail.com>:
>>>
>>>> Hi All
>>>> I am observing a strange issue on my storm cluster (6 nodes). After
>>>> submitting the topology I can see some 3-7 mins time when worker slot got
>>>> occupied.
>>>> I am clueless why it is taking so long to occupy the worker slot.I also
>>>> observe the same with the work-count topology so it is not topology
>>>> specific.
>>>> Could somebody provide me some pointers so that I can identify the
>>>> problem.
>>>>
>>>> Thanks
>>>>
>>>
>>>
>>
>
--
Regards,
Abhishek Agarwal
Re: Storm topology taking too long to get assigned on worker slot
Posted by Sanjeev Verma <sa...@gmail.com>.
looking for some pointers so that can pin point the problem.
Thanks
On Tue, May 24, 2016 at 6:09 PM, Sanjeev Verma <sa...@gmail.com>
wrote:
> its 10.0.2.4
>
> On Tue, May 24, 2016 at 9:45 AM, 马哲超 <ma...@gmail.com> wrote:
>
>> Which version do you use?
>>
>> 2016-05-23 19:48 GMT+08:00 Sanjeev Verma <sa...@gmail.com>:
>>
>>> Hi All
>>> I am observing a strange issue on my storm cluster (6 nodes). After
>>> submitting the topology I can see some 3-7 mins time when worker slot got
>>> occupied.
>>> I am clueless why it is taking so long to occupy the worker slot.I also
>>> observe the same with the work-count topology so it is not topology
>>> specific.
>>> Could somebody provide me some pointers so that I can identify the
>>> problem.
>>>
>>> Thanks
>>>
>>
>>
>
Re: Storm topology taking too long to get assigned on worker slot
Posted by Sanjeev Verma <sa...@gmail.com>.
looking for some pointers so that can pin point the problem.
Thanks
On Tue, May 24, 2016 at 6:09 PM, Sanjeev Verma <sa...@gmail.com>
wrote:
> its 10.0.2.4
>
> On Tue, May 24, 2016 at 9:45 AM, 马哲超 <ma...@gmail.com> wrote:
>
>> Which version do you use?
>>
>> 2016-05-23 19:48 GMT+08:00 Sanjeev Verma <sa...@gmail.com>:
>>
>>> Hi All
>>> I am observing a strange issue on my storm cluster (6 nodes). After
>>> submitting the topology I can see some 3-7 mins time when worker slot got
>>> occupied.
>>> I am clueless why it is taking so long to occupy the worker slot.I also
>>> observe the same with the work-count topology so it is not topology
>>> specific.
>>> Could somebody provide me some pointers so that I can identify the
>>> problem.
>>>
>>> Thanks
>>>
>>
>>
>
Re: Storm topology taking too long to get assigned on worker slot
Posted by Sanjeev Verma <sa...@gmail.com>.
its 10.0.2.4
On Tue, May 24, 2016 at 9:45 AM, 马哲超 <ma...@gmail.com> wrote:
> Which version do you use?
>
> 2016-05-23 19:48 GMT+08:00 Sanjeev Verma <sa...@gmail.com>:
>
>> Hi All
>> I am observing a strange issue on my storm cluster (6 nodes). After
>> submitting the topology I can see some 3-7 mins time when worker slot got
>> occupied.
>> I am clueless why it is taking so long to occupy the worker slot.I also
>> observe the same with the work-count topology so it is not topology
>> specific.
>> Could somebody provide me some pointers so that I can identify the
>> problem.
>>
>> Thanks
>>
>
>
Re: Storm topology taking too long to get assigned on worker slot
Posted by 马哲超 <ma...@gmail.com>.
Which version do you use?
2016-05-23 19:48 GMT+08:00 Sanjeev Verma <sa...@gmail.com>:
> Hi All
> I am observing a strange issue on my storm cluster (6 nodes). After
> submitting the topology I can see some 3-7 mins time when worker slot got
> occupied.
> I am clueless why it is taking so long to occupy the worker slot.I also
> observe the same with the work-count topology so it is not topology
> specific.
> Could somebody provide me some pointers so that I can identify the problem.
>
> Thanks
>