You are not convinced that the .BannedIP package actually contains what you expect?
You can programmatically use statistics to find out what is going on. Given that you have a configuration bit, Ehcache should never fail (as long as you have less than 80k instances of BannedIP).
In addition, to understand what this means: you are in the process of updating BannedIP instances, but when the commit occurs, the SoftLock that Hibernate set to mark this entry is in the process of being updated, for some reason evicted ... Interestingly, Why are these updates taking so long?
Also, although Iām just thinking out loud, could it be some other process (mass update) that will invalidate the entire cache region?
source share