Just had to workaround a bug that had me pulling my hair out for most of the last month.
Part of the problem was isolating exactly which combination of server/loadbalancer/cluster/software was causing the issues. As a quick check, does anyone else run Roller 4.0.x on a Glassfish cluster (I stress that's *cluster*, the problem only seems to manifest if you use in-memory session replication)? If not, I'll have to put it on the list of 'things to doublecheck when I have spare time', and we all know how often they get looked at...
