I know that Paul Ferraro mentioned this on IRC a while back. However he wasn't able to produce a case or trace when it occurred. It seems like a bug to me though, do you have a test case to reproduce this or a trace of it occurring? With those we should be able to remedy the issue for you guys.
It would be very helpful if you could provide:
a) event.toString() (so that we can see if it was pre- or post-event etc.)
b) your cache configuration
c) are you using any flags when accessing the cache?
I don't have a specific scenario where it happens but we see it from time to time while removing entries from a replicated cache.
It happens on a post-event.
Here's the cache configuration that we are using :
<stack-file name="external-file" path="jgroups.xml"/>
<transport stack="external-file" />
<distributed-cache name="session" mode="ASYNC" owners="2">
<replicated-cache name="global" mode="ASYNC">
We do not use any specific flags when accessing the cache.
In our testing, we run two instances of the application on the same machine.