[
https://issues.apache.org/jira/browse/HIVE-12951?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15605111#comment-15605111
]
Rui Li commented on HIVE-12951:
-------------------------------
Hi [~richard_xin], pre-warm is intended to wait until some containers are
launched before we decide the parallelism/start the job. In general, it helps
us to have a higher parallelism and better task scheduling. However it doesn't
remove the start-up overhead for the 1st query - anyway we have to wait for the
containers to be launched to run the query.
[~xuefuz] please add if I missed anything.
> Reduce Spark executor prewarm timeout to 5s
> -------------------------------------------
>
> Key: HIVE-12951
> URL: https://issues.apache.org/jira/browse/HIVE-12951
> Project: Hive
> Issue Type: Bug
> Components: Spark
> Affects Versions: 1.2.0
> Reporter: Xuefu Zhang
> Assignee: Xuefu Zhang
> Fix For: 2.1.0
>
> Attachments: HIVE-12951.patch, HIVE-12951.patch
>
>
> Currently it's set to 30s, which tends to be longer than needed. Reduce it to
> 5s, only considering jvm startup time. (Eventually, we may want to make this
> configurable.)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
