There is a lot of contention in the Bandana table that can cause deadlocks, which can lead to Confluence running out of database connections and falling over with messages like:

2011-05-04 05:17:50,594 ERROR [TP-Processor1545] [sf.hibernate.util.JDBCExceptionReporter] logExceptions Cannot get a connection, pool error Timeout waiting for idle object

Things to do:

• Modify the bandanacontext and bandanakey columns adding not null constraints
• Add a unique constraint to the Bandana table on the bandanacontext and bandanakey columns
• Modify the HibernateConfluenceBandanaRecordDao.saveOrUpdate method so that if an insert fails with a violation of the constraint, it tries an update
• Remove the synchronizing lock from the HibernateConfluenceBandanaRecordDao.saveOrUpdate method.