You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@impala.apache.org by "Tim Armstrong (Jira)" <ji...@apache.org> on 2020/01/29 01:48:00 UTC

[jira] [Resolved] (IMPALA-4400) Aggregate runtime filters locally

     [ https://issues.apache.org/jira/browse/IMPALA-4400?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Tim Armstrong resolved IMPALA-4400.
-----------------------------------
    Fix Version/s: Impala 3.4.0
       Resolution: Fixed

> Aggregate runtime filters locally
> ---------------------------------
>
>                 Key: IMPALA-4400
>                 URL: https://issues.apache.org/jira/browse/IMPALA-4400
>             Project: IMPALA
>          Issue Type: Improvement
>          Components: Backend
>    Affects Versions: Impala 2.8.0
>            Reporter: Marcel Kinard
>            Assignee: Tim Armstrong
>            Priority: Major
>              Labels: multithreading
>             Fix For: Impala 3.4.0
>
>
> At the moment, runtime filters are sent from each fragment instance directly to the coordinator for aggregation (ORing) at the coordinator.
> With multi-threaded execution, we will have an order of magnitude more fragment instances per node, at which point the coordinator would become a bottleneck during the aggregation process. To avoid that, we need to aggregate the local instances' runtime filters at each node before sending the filter off to the coordinator.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)