storm-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Anuj Kumar <anujs...@gmail.com>
Subject Re: Whole topology restarts when a worker crash
Date Tue, 01 Apr 2014 14:24:40 GMT
Hi Quentin,

What is your Storm version? I am running on 0.8.2.

Regards,
Anuj


On Tue, Apr 1, 2014 at 7:26 PM, Quentin de G. <quentin.de.gr@gmail.com>wrote:

> Hey Anuj, thanks for your answer.
> Unfortunately that's not what I'm seeing in Storm UI.
>
> I have 2 machines running the workers, with one worker / machine.
> When I kill a worker (kill -9 pid), even the tasks that were not running
> on this worker are restarted (uptime of all tasks reset).
>
>
> Maybe I'm doing something wrong in my topology, or some config parameters
> are wrong.
>
> Thanks,
>
> Quentin
>
>
>
>
> On Tue, Apr 1, 2014 at 3:26 PM, Anuj Kumar <anujsays@gmail.com> wrote:
>
>> Hi Quentin,
>>
>> If the worker crashes only the associated executor threads running the
>> tasks are restarted. Not the entire topology.
>>
>> Regards,
>> Anuj
>>
>>
>> On Tue, Apr 1, 2014 at 6:44 PM, Quentin de G. <quentin.de.gr@gmail.com>wrote:
>>
>>> Hello everyone,
>>>
>>> It seems that when one worker crashes, the whole topology is restarted.
>>> Is there a way to tune that, like only restarting tasks that were bound
>>> to this worker?
>>>
>>> Thanks.
>>> --
>>>    Quentin de Grandmaison
>>>
>>>  / LinkedIn<http://www.linkedin.com/profile/view?id=135068230&trk=tab_pro>profile
/ Profil
>>> Viadeo <http://www.viadeo.com/profile/002e4ak5nhwbe2h> Mon CV<http://labki.net/>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>
>
>
> --
>    Quentin de Grandmaison
>
>  / LinkedIn<http://www.linkedin.com/profile/view?id=135068230&trk=tab_pro>profile
/ Profil
> Viadeo <http://www.viadeo.com/profile/002e4ak5nhwbe2h> Mon CV<http://labki.net/>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>

Mime
View raw message