We aren't talking about asynchrously replicating real updates - those still replicate synchronously. But if #2 is the replication of another cache fill, then it simply doesn't matter when it happens. (Unless it overwrites a real update.) The asynchronous replicating of putFailFast has been in the codebase as far back as May 05, which is as far as the CVS history extends. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3977574#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...