Well, the cluster has restarted. The only strange thing I can see is that node 2 is used less than the others (1%) collie node info Id Size Used Use% 0 216 GB 21 GB 9% 1 216 GB 20 GB 9% 2 216 GB 2.6 GB 1% 3 112 GB 11 GB 10% Total 760 GB 55 GB 7%