[
https://issues.apache.org/jira/browse/HIVE-16893?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
anishek updated HIVE-16893:
---------------------------
Attachment: HIVE-16893.3.patch
fixed sankar's comments
> move replication dump related work in semantic analysis phase to execution
> phase using a task
> ---------------------------------------------------------------------------------------------
>
> Key: HIVE-16893
> URL: https://issues.apache.org/jira/browse/HIVE-16893
> Project: Hive
> Issue Type: Sub-task
> Components: HiveServer2
> Affects Versions: 3.0.0
> Reporter: anishek
> Assignee: anishek
> Fix For: 3.0.0
>
> Attachments: HIVE-16893.2.patch, HIVE-16893.3.patch
>
>
> Since we run in to the possibility of creating a large number tasks during
> replication bootstrap dump
> * we may not be able to hold all of them in memory for really large
> databases, which might not hold true once we complete HIVE-16892
> * Also a compile time lock is taken such that only one query is run in this
> phase which in replication bootstrap scenario is going to be a very long
> running task and hence moving it to execution phase will limit the lock
> period in compile phase.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
