Automated purging of data
By default Oracle SOA Suite leaves into the database (dehydration store) all the instances that have been executed successfully or failed or aborted for one or the other reason. This means that over time the database grows and in a busy production environment sooner or later you will run out of allocated disk space. The natural solution is to allocate more space to the database which ultimately leads to a situation where database if full and the disks have no free space and then you are really starting to have serious fun.
In order to avoid this, I strongly recommended to schedule the purge scripts of SOA Suite. You can see more instructions here:
It’s also a best practice to delete the archive logs of database (see Backing Up and Deleting Archived Redo Log Files in Backup and Recovery Reference).
Found via dear friend Mr. Google Marc Kelderman's SOA blog and the following resources on improving purging performance there. If you have issues with your purging strategy, you could take a look at these resources as they seem very good pointers: