Replacing a failed cluster node?

[Lebha]

Hello!

We're running a single-node cluster and the passive node started getting kernel panics, probably due to failing hardware. As we probably need to replace the whole machine, I'm wondering if anybody has replaced a failed cluster node?

The normal cluster installation instructions proceed with steps executed taking turns on both the active and passive nodes. How does one build just one cluster node from scratch?

[Klug]

Take the harddrives out of the failing node, replace the node, put the HD in 8)