[
https://issues.apache.org/jira/browse/HIVE-10982?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14906096#comment-14906096
]
Vaibhav Gumashta commented on HIVE-10982:
-----------------------------------------
[~libing] Thanks for the patch. I'll review soon. I have one high level
question: with this patch the client can specify an arbitrary large value for
the fetch size and in a scenario with many concurrent clients fetching results,
it might bring down HS2 with OOM. I think we should have a configurable upper
limit on the server side as well.
> Customizable the value of java.sql.statement.setFetchSize in Hive JDBC Driver
> ------------------------------------------------------------------------------
>
> Key: HIVE-10982
> URL: https://issues.apache.org/jira/browse/HIVE-10982
> Project: Hive
> Issue Type: Improvement
> Components: JDBC
> Affects Versions: 1.2.0, 1.2.1
> Reporter: Bing Li
> Assignee: Bing Li
> Priority: Critical
> Attachments: HIVE-10982.1.patch
>
>
> The current JDBC driver for Hive hard-code the value of setFetchSize to 50,
> which will be a bottleneck for performance.
> Pentaho filed this issue as http://jira.pentaho.com/browse/PDI-11511, whose
> status is open.
> Also it has discussion in
> http://forums.pentaho.com/showthread.php?158381-Hive-JDBC-Query-too-slow-too-many-fetches-after-query-execution-Kettle-Xform
> http://mail-archives.apache.org/mod_mbox/hive-user/201307.mbox/%3ccacq46vevgrfqg5rwxnr1psgyz7dcf07mvlo8mm2qit3anm1...@mail.gmail.com%3E
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
