I have a query saying "select count(*) as y0_ from Service this_ where (this_.configstate=? and 1=1)" and another similar query saying "select count(*) as y0_ from Service this_ where (this_.configstate=? and 1=1 and this_.state like ?)" running from one JVM . Service table has been updated through some other JVM with one entry deletion and one entry addition simultaneously. Now, the first query returns stale data while the other query return appropriate data. The only other difference in my eyes is that the first query is very frequently used while the other isn't.

This issue gets corrected post two hours. What could be the probable reason for this issue?

You don't have a @Cache annotation on your entity do you?

No, the cache is set as NoCacheProvider and session is being opened and closed everytime pre and post queries.

Are you using Master/Slave replication and there is a lag between Master and Slave?

There is no master slave as well.. Just that database is accessed from two JVMs as explained. But the two queries I posted are through same JVM.