Title: grasping at straws.

Hi folks - I'm trying to track down a new memory/socket/filehandle leak that occurred in our product.  I've narrowed it down to 1 of 3 things that changed for this release.   One of them is hsqldb.   Things to know though:

OS = Solaris 8
Java = 1.4.2_08
HSQLDB = 1.8rc9
Mode = Server = connect to 2 separate hsqldb instances on different ports.

The test to fail takes over 12 hours so our last test succeeded when we changed out the 3 items - one of the 3 was reverting hsqldb back to 1.7.2  - since the test succeeded I'm now running again with just hsqldb put back to version 1.8 to see if it causes us to fail.   I don't see how it could be hsqldb but we open a long running connection  with the db that lasts during the entire test.   Prepared statements are used.   Queries and updates are happening throughout.   Now my question to you is did anything change on the DRIVER side since the hsqldb processes aren't running out of memory/threads/sockets it's the client of the database that is?    I'm sure I'll find its one of the other 2 things that's causing the memory leak but I'd like to rule out hsqldb if I can.

 

Andrew Chandler

Manager DataCollection Group     


Visionael Corporation             Phone : (918) 388-8308

9717 E. 42nd Street               Fax   : (918) 663-1456 <mailto:[EMAIL PROTECTED]>

Suite 200                         Email : [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>
Tulsa, OK 74146-3680              WEB   : www.visionael.com <http://www.visionael.com/>

Reply via email to