It seems that the default behavior of drbd is to reboot the primary node of a cluster if it gets a split-brain. For a 3 node cluster this might make sense. For a 2 node cluster it means that if the secondary fails then the primary goes down too. How do I configure drbd to not do this and are there any issues with doing so? -- My Main Blog http://etbe.coker.com.au/ My Documents Blog http://doc.coker.com.au/