On 07/10/2012 08:35 PM, dj wrote: > Greetings! > > I am new to setting up Linux Clusters. I am setting up a two node HA > cluster (RHEL 5.8) for Postgres. We like to have NFS > mount fail over from node 1 to 2. > > My understanding is that DRBD needs shared DISK & we don't have that. How > can we set up the HA with Hearbeat > for NFS with out DRBD? > > I am using :heartbeat-2.1.3-3.el5.centos.x86_64 > > any pointers or suggestions would be very much appreciated. > > Thank you > Mike
You have a few topics here; 1. DRBD is, in concept, like traditional mirroring / RAID 1. You take a chunk of storage on each node and tie them together to get a common storage. To continue the analogy; RAID 1: /dev/sda5 + /dev/sdb5 == /dev/md0 DRBD: node1:/dev/sda5 + node2:/dev/sda5 == /dev/drbd0 Both nodes would see the '/dev/drbd0' device and anything written to that device would be synchronously copied to the other node, so '/dev/drbd0' would always look the same on both nodes. Now, that is just a block device. As with a traditional raid array, you need to put a filesystem on top of it to actually use it (unless you use the raw space the back a VM, but that's another topic). If you want simple failover, then you can put a normal FS like ext4 on the DRBD device and only mount it on the one node (the most common setup). If the primary node fails, then you would mount the /dev/drbd0 device on the backup and away you go. There is more to it, of course, but that should give you the basic idea. 2. If it's a new cluster, use RHEL (CentOS) 6.3, not 5.8. It will be supported 3 years longer and it the focus of most of the development and testing these days. 3. The heartbeat project is deprecated and it's use is *strongly* discouraged. You should use either corosync + pacemaker or, if you want RHEL support, corosync + rgmanager. Digimer -- Digimer Papers and Projects: https://alteeve.com _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
