[
https://issues.apache.org/jira/browse/KAFKA-190?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jay Kreps resolved KAFKA-190.
-----------------------------
Resolution: Won't Fix
I think this is less relevant with the new fsync management and replication
code.
> Parallelize time-based flushes
> ------------------------------
>
> Key: KAFKA-190
> URL: https://issues.apache.org/jira/browse/KAFKA-190
> Project: Kafka
> Issue Type: New Feature
> Reporter: Jay Kreps
>
> Currently all time based flushes are done sequentially. If the time-based
> flush is the only policy that is used this may limit throughput as a single
> thread may not saturate all the disks. It would be good to make the number of
> flush threads be configurable. We can probably default it to two threads
> (threads are cheap).
--
This message was sent by Atlassian JIRA
(v6.2#6252)
