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

Posted to issues@flink.apache.org by "Márton Balassi (JIRA)" <ji...@apache.org> on 2014/12/12 19:34:13 UTC

[jira] [Created] (FLINK-1323) Implement memory management profiles

Márton Balassi created FLINK-1323:
-------------------------------------

             Summary: Implement memory management profiles
                 Key: FLINK-1323
                 URL: https://issues.apache.org/jira/browse/FLINK-1323
             Project: Flink
          Issue Type: Improvement
          Components: Distributed Runtime
    Affects Versions: 0.8-incubating
            Reporter: Márton Balassi


Flink currently allocates a sizeable chunk of the JVM heap for the framework, where hashing and sorting is done amongst others. This enables graceful spilling to disk, when this datasize exceeds memory, however might be counterproductive for jobs with large user state and is completely unutilized by streaming at the moment.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)