> Heh, running into that right now :( Currently gluster is running checksums on 12 > VM's (500GB data), pegging the CPU on all nodes. This after just rebooting one > node. > > Ceph would perform better in this case? yes (much better).