Dirty data -- the bean is not always king

[kellymac]

Here is the problem,

The database has points of contact other than any given bean, scheduled jobs can change data, triggers in the database can modify data while it is stored, and another user's bean can change stuff. The bean loads the data and thinks that it is the only one that can touch it. I have a way to tell that the data is dirty when storing but by then it is too late. Some process thought everything was fine and went on and now the data can’t be stored. There is also a chance that the data was changed by another process after the load and it is already dirty before anything is even served up to the client.

Is there a way to get the container to check and see if the data is dirty and re load it or force the container to always do a load?

[marc.fleury]

you want to use option B for reload or D for time expiration of the cache

[kellymac]

Thanks for the reply!

We tried to do this but were not successful. I am sure it is documented but it would be so helpful if you could explain exactly how to make this change.

Thank you,

KellyMac

[kellymac]

Thanks for the reply!

We tried to do this but were not successful. I am sure it is documented but it would be so helpful if you could explain exactly how to make this change.

Thank you,

KellyMac

[juha]

conf/standardjboss.xml

[kellymac]

Well,

I got it working. I made entries in my project’s jboss.xml file. I added container configuration entries for each of the standard container configurations that were not already set to ‘B’ in standardjboss.xml.

My entries just have the commit-option tag. When you are not specifying a container it is not always clear which one jboss is using so I just changed them all. Do you see a problem with this approach?

Thanks again for your help, Things seem much better now.


KellyMac