I see where you got the idea about eviction from. :-) I assume your cache is configured for INVALIDATION_SYNC or INVALIDATION_ASYNC. Seems in 1.x the cluster wide invalidation was implemented by telling the other nodes to do an evict. And that means the call uses the 0 ms timeout meant for regular evictions. Note sure what advice I can give you here, other than to try JBC 2.x (2.1.1.GA is latest stable release). JBC 2.x handles invalidation differently. Perhaps someone else on this forum knows a workaround. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4157484#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...