Just personal experience, sorry. * it deadlocks. Often. * it often doesn't recover from network partitions. Symptoms include both sides of the partition nev

Just personal experience, sorry. * it deadlocks. Often. * it often doesn't recover from network partitions. Symptoms include both sides of the partition never merging, and deadlocks. Recovering requires rebooting the entire cluster at once. Due to their locking strategy, a rolling reboot of the cluster isn't enough. *some of the developers didn't seem to demonstrate understanding of races, or distributed systems. When I reported racing bugs, they asked if I could attach reliable unit tests.