We run ehcache 1.5 in a cluster with RMI replication and update servers one by one (think of a load balancer / proxy on top and zero upgrade).
We don't care about serialVersionUID . The problem is that very bad things (up to disconnection) can happen if you have two versions of an object in a replicated cache. That is, if one of the servers running the old code replicates the item to the new server, where its class has been changed.
We usually get around this by launching a new cluster for upgraded servers on a different port, but it's pretty ugly and fragile.
So, the question is: is clustered, replicated ehcache properly complying with serialVersionUID ? That is, is object replication not trying if the version of the local class is different?
Thanks for the intuitive guesses, but I am looking for the hardest data possible, preferring official documents.
source share