This week's book giveaway is in the Agile and Other Processes forum.
We're giving away four copies of The Little Book of Impediments (e-book only) and have Tom Perry on-line!
See this thread for details.
Win a copy of The Little Book of Impediments (e-book only) this week in the Agile and Other Processes forum!
  • Post Reply
  • Bookmark Topic Watch Topic
  • New Topic

ecperf problems

 
Scott Nyen
Greenhorn
Posts: 3
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
I have installed the ecperf benchmark on a Websphere 5.0 server accesing a host DB2 database. When I attempt to execute load against this server I get thousands "SQL0913N Unsuccessful execution caused by deadlock or timeout. Reason code "00C9008E" errors. I am using COM.ibm.db2.jdbc.DB2XADataSource as my datasource implemnation.
This all runs fine in the web demo application. I have tryed implementing it as J2EE 1.2 and 1.3 and have the same problems.
Anyone have any ideas?
 
Scott Nyen
Greenhorn
Posts: 3
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
Also... These are deadlocks, not timeouts. The Database guys did a trace and confirmed this...
 
Thomas Taeger
Ranch Hand
Posts: 311
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
Originally posted by Scott Nyen:
...deadlocks ...

Hi Scott,
nobody can guess where the deadlock (or even deadlocks) comes from without debugging or tracing your code.
A deadlock just means two objects waiting for each other releasing an object or resource the other one needs. Therefore behaviour may very well differ when running in a test server having another (slower...) timing. But NOT running into a deadlock with the test server just might mean that you had been "lucky". It does not guarantee that there are no deadlock conditions.
Even debugging (if possible at all) often does not detect the deadlock.
You might need to add short tracing output like "entering xyz ..." and "leaving xyz" to find the pair that entered but never left that method.
Then read in the code what ressources both need. These are candidates for deadlocks.
Hope that helps a little bit. I know how deadlocks can drive one creasy...
But be aware that even just adding trace output can "fix" the problem - or defer it to another spot of code ...
Thomas.
 
Scott Nyen
Greenhorn
Posts: 3
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
I was hoping someone had experianced this before. ecperf is a standard benchmark so it shouln't be a code problem. I was thinking there maybe a deployment descriptor or other setting that needs to be changed on Websphere or perhaps a driver issue...
 
  • Post Reply
  • Bookmark Topic Watch Topic
  • New Topic