On 14/07/13 20:33, Sergio Alex wrote:
Hi, I need to create a real-time replication of a partition between two servers with centos 6.4, I am researching and finding difficulty with drbd, corosync, someone indicate material, do not need to run script automatically if a node drops only replicate. Grateful.
What problem are you having with DRBD? I use DRBD on CentOS 6 with corosync + cman + rgmanager. If you need simple fail-over, it's easy. If you want concurrent access from both nodes, you'll need a cluster-aware filesystem like GFS2.
The only real trick is to be sure you use fencing. This prevents both nodes from trying to use the storage when they think the other node is dead (a condition called a "split-brain").
I have a tutorial that covers all of this. It's designed for running VMs on DRBD, but you can ignore the VM part;
https://alteeve.ca/w/2-Node_Red_Hat_KVM_Cluster_Tutorial
digimer