[
https://issues.apache.org/jira/browse/SPARK-4873?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14249716#comment-14249716
]
Apache Spark commented on SPARK-4873:
-------------------------------------
User 'zsxwing' has created a pull request for this issue:
https://github.com/apache/spark/pull/3721
> WriteAheadLogBasedBlockHandler improvement
> ------------------------------------------
>
> Key: SPARK-4873
> URL: https://issues.apache.org/jira/browse/SPARK-4873
> Project: Spark
> Issue Type: Improvement
> Components: Streaming
> Reporter: Shixiong Zhu
> Priority: Minor
>
> 1. Move the block store work to a separate thread pool. IO work usually runs
> longer than CPU work. Running IO work in `Implicits.global` may block other
> work that want to run in `Implicits.global`.
> 2. Use `Future.zip` instead of `Future.flatMap`(for-loop). `zip` implies
> these two Futures will run concurrently, while `flatMap` usually means one
> Future depends on the other one.
> 3. Fix a unit test in `ReceivedBlockTrackerSuite` which may fail occasionally.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
