accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Roshan Punnoose (JIRA)" <>
Subject [jira] [Commented] (ACCUMULO-1224) Custom Thread Pool in BatchScanner
Date Fri, 29 Mar 2013 20:49:15 GMT


Roshan Punnoose commented on ACCUMULO-1224:

Maybe the right answer is to have a Factory definition for creating threads that can be backed
by a thread pool. It would implement isOpen and close methods. And any external implementations,
even if they wrap a thread pool, would have to decide how to "close", whether that means to
do a shutdownNow, or wait for threads to finish, etc. 

Where would a user specify the BatchConfig? In the Instance when it starts? Maybe it is just
a Hadoop Configuration object that is wrapped with methods to add the thread pool factory
class name, etc. 

Btw, this is probably not the right place to ask this, but how come the Connector class is
a class and not an interface? Seems like there was probably a reason for this? (Instance too)
> Custom Thread Pool in BatchScanner
> ----------------------------------
>                 Key: ACCUMULO-1224
>                 URL:
>             Project: Accumulo
>          Issue Type: Improvement
>          Components: client
>            Reporter: Roshan Punnoose
>            Assignee: Billie Rinaldi
>            Priority: Minor
>         Attachments: threadPool.patch
> I would like the ability to supply my own ThreadPool(ExecutorService) when the BatchScanner
is created. This would give me the ability to throttle how many threads are created in an

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see:

View raw message