|
I'm having the same issue, also on 10g. It happens when the data being updated is large-ish, i.e., >1k or so in VARCHAR2 columns in batch size >1. This isn't scientific, just something I've noticed.
When the problem occurs, I notice that the system lots many unrelated tables (though none blocking in my case). I'm still researching -- more to come as I find it. Setting hibernate.jdbc.batch_size=0 alleviates the problem.
Here's a little diagnostic info from my setup:
08:21:43,642 INFO [Environment] Hibernate 3.0.5
08:21:46,486 INFO [SettingsFactory] RDBMS: Oracle, version: Oracle Database 10g Enterprise Edition Release 10.1.0.4.0 - Production
With the Partitioning, OLAP and Data Mining options
08:21:46,486 INFO [SettingsFactory] JDBC driver: Oracle JDBC driver, version: 10.2.0.1.0
08:21:46,533 INFO [Dialect] Using dialect: org.hibernate.dialect.Oracle9Dialect
It is noteworthy that this happens to me on relatively new code -- I can reproduce it in my test environments but not in production. I do not know yet if this is a code problem or a difference in database config. Our prod database (not running new code yet) has different memory settings (larger SGA) so this may be a factor.
I'll be happy to collaborate on this with you offline so we don't spam the board with our discussion.
-Chris
|