Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
I am setting up a test drbd 9 cluster with Pacemaker. The documentation states that you can either use drbd as if it is a SAN or have Pacemaker control drbd manually. I am using drbdmanage to handle the resource creation and was hoping to keep everything as simple as possible and have drbdmanage do most of the hard work and also allow automatic resource promotion. At this stage, I intend to set up just an active/passive cluster. So far, I have zfs and drbd set up as I want. I am using a zfs -> drbd -> zfs -> nfs configuration. So, for example, I have a drbd device called vm-data-01 and I have created a zpool called vm-data-01 and on my first cluster this has been exported via nfs. It works fine. I can also migrate this resource to my second node and again, it works fine. The nfs share also exports correctly. My main question is whether or not this is the "correct" way to do it and whether there are any serious pitfalls. Also, how do I prevent auto-promotion of the drbd block device if the node I am migrating to is not up to date? Many thanks. Martyn