[
https://issues.apache.org/jira/browse/PHOENIX-544?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
James Taylor resolved PHOENIX-544.
----------------------------------
Resolution: Duplicate
Duplicate of PHOENIX-36
> Specify desired parallelization concurrency as a percentage of available
> resources
> ----------------------------------------------------------------------------------
>
> Key: PHOENIX-544
> URL: https://issues.apache.org/jira/browse/PHOENIX-544
> Project: Phoenix
> Issue Type: Task
> Reporter: James Taylor
>
> Instead of configuring a fixed size number of threads (maxConcurrency and
> targetConcurrency), @larsh had a good idea to specify it as a percentage of
> available server-side resources. So the number of chunks to split up a scan
> and run them in parallel would become:
> parallelizationFactor * numberOfRegionServers
> We could potentially do even better by returning in our compatibility checker
> (that runs once at init time per cluster) the number of handler threads for
> each region server. Then we'd have a more accurate number of total available
> resources.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
