I am using Hibernate 2.1.4 and I am noticing a weird behavior from something that I believe I must have done wrong. As I understand session.find should return results directly from the DB and not from the cache. I am returning a list of results by the following:

Iterator iter = session.find("from com.scottrade.nagios.hibernate.Hoststatus host").iterator();

I am using that instead of session.iterate because I do not want to use the cache. The reason is that the database is being written to by another application and I want to read in the fresh data. In this case, that is not occurring.

Is there something that I am missing?

Thank you,

Andrew Janian
OMS Developer
Scottrade Financial Services
ajanian@scottrade.com

Did you change the FlushMode to FlushMode.COMMIT? Because then it is possible that you receive stale data. In this case, a flush to the database is done when you commit the transaction and not before you call the find() method.

Reto

you may use session.clear() before calling find or use a new session

your session should have a short life.... if you are sometimes seeing stale data, it may be because you are considering the session as a global cache --> it isn't!

The other application is not using hibernate. That is the root of the problem. I am not able to use session.clear() because that makes the query take 8 seconds per use. This page is supposed to refresh every 10 seconds, so that would not make sense.

This is not really a great place to use hibernate, except for the OR mappings. I don't want the cache at all, but I seem to paying to much in time for the OR mappings.

As a last attempt, how do I set the cache expiration for an object?

Thanks for your help,

Andrew

why?

when you do it in pure sql, how long does it take?

are you sure you need an orm?

are you sure tunning the query is not the real problem?

I have the same problem. When a database is accessed from another application we get stale data. To overcome this problem we need to refresh each list item. But why? The Session.find method already has read data by a query like "select * from ...". But only primary keys are controlled.

I thought that was the precise difference between find and iterate. Are we misunderstanding something?