You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Nikolay Izhikov (Jira)" <ji...@apache.org> on 2020/03/26 14:56:00 UTC
[jira] [Updated] (IGNITE-12826) Poor JDBC Cache Store performance
due to default fetch size
[ https://issues.apache.org/jira/browse/IGNITE-12826?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Nikolay Izhikov updated IGNITE-12826:
-------------------------------------
Fix Version/s: 2.9
> Poor JDBC Cache Store performance due to default fetch size
> -----------------------------------------------------------
>
> Key: IGNITE-12826
> URL: https://issues.apache.org/jira/browse/IGNITE-12826
> Project: Ignite
> Issue Type: Bug
> Components: cache
> Affects Versions: 2.8
> Reporter: Alexey Kukushkin
> Assignee: Nikolay Izhikov
> Priority: Major
> Labels: sbcf
> Fix For: 2.9
>
> Attachments: ignite-12826-vs-2.8.patch
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> JDBC "fetchSize" parameter specifies the number of rows to be fetched from the database when additional rows are needed. For most drivers it is 10 by default. Larger fetchSize can significantly improve performance due to less network roundtrips (at expense of greater memory consumption).
> For some reason out-of-box JDBC POJO Cache Store uses default fetchSize in the loadCache method implementation.
> We have very poor loadCache performance when loading large amount of data from Oracle with the default fetchSize of 10. We tried setting fetchSize to 20K and that improved performance 40 times.
> We need to use JdbcDialect#fetchSize in the loadCache implementation so that users could implement a custom JdbcDialect to configure fetchSIze.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)