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

Posted to jira@kafka.apache.org by "Jason Gustafson (Jira)" <ji...@apache.org> on 2019/08/28 00:12:00 UTC

[jira] [Created] (KAFKA-8841) Optimize Partition.maybeIncrementLeaderHW

Jason Gustafson created KAFKA-8841:
--------------------------------------

             Summary: Optimize Partition.maybeIncrementLeaderHW
                 Key: KAFKA-8841
                 URL: https://issues.apache.org/jira/browse/KAFKA-8841
             Project: Kafka
          Issue Type: Improvement
            Reporter: Jason Gustafson
            Assignee: Jose Armando Garcia Sancio


We've observed a noticeable cost in `Partition.maybeIncrementLeaderHW` during profiling. Basically just all the unnecessary collection copies and iterations over the replica set. Since this is on the hot path for handling follower fetches, we should probably put some effort into optimizing this. 

A couple ideas:
 # Don't do all those copies.
 # Checking whether hw needs incrementing is not necessary unless an end offset actually changed



--
This message was sent by Atlassian Jira
(v8.3.2#803003)