Thanks, Michael.
From: eclipselink-users-bounces@xxxxxxxxxxx [mailto:eclipselink-users-bounces@xxxxxxxxxxx]
On Behalf Of Michael O'Brien
Sent: Friday, May 11, 2012 2:42 PM
To: EclipseLink User Discussions
Subject: Re: [eclipselink-users] Eclipselink & Microsoft SQL Server
Sri,
Hi, I have used EclipseLink 2.3.2 as a JPA 2.0 provider with Sybase ASE 15.7 using the JConnect7 JDBC 4 driver with no issues.
EclipseLink and Sybase in my case work fine.
The only issues I has was with bulk import – however this was with Sybase itself. A brute force 100K row import ran out of locks. However
when I used proper batch writing parameters I was able to maintain a sustained import rate of 200 rows/sec (in my case).
Note: Most of the issues I encounter are due to lack of properly configured JPA or JDK parameters on my part. For example my heap
when it passes 70% of the max using a non-JRockit or non[-server] flagged SUN JRE will encounter up to a 900% slowdown due to excessive GC activity. In this case I went from a 73 min import to an 8 min import by changing to the following [-Xmx1024M –server]
when my heap hovered around 480.

Sybase and SQL Server have a common ancestor version so I would expect similar behavior from Microsoft’s version.
/michael
Have you used Eclipselink to interact with a Microsoft SQL server database? Are there any issues to be aware of?
Are there any minimum database version requirements? How about Eclipselink version?
I am starting work on an application whose reference data source is available in a SQL server database.
If I find that the Eclipselink <<>> SQL Server nexus is not as robust as working with Oracle I will plan to use Oracle for the application’s operational store.