Error Message:

The cluster database on the local node is in an invalid state. Please start another node before starting this node.

Explanation:

Whenever a quorum device is changed to a new resource, a tombstone file is left behind on the old quorum resource. This error can occur when the quorum device has been changed from one drive to another. In particular this error can occur if the quorum device was changed when one of the nodes in the cluster was down. For example, while node A was not running, on node B, the quorum device was changed to a different resource.\r\nAs a result, node A has outdated quorum-device information, and node B has the current quorum-device information. If node B now goes down and node A comes up, it tries to get access to the old quorum device. Before it does that it checks to see whether there is a "tombstone" file. The tombstone file doesn't say where the new quorum resource is; it just says that what was the quorum device on node A no longer is the quorum device. The result is that node A can only join a cluster. The presence of a quorum tombstone marks the end of the use of a resource as a quorum device.

User Action:

Strongly recommended fix: Restart node B and have node A join the cluster. In the worst-case scenario, in which node B is really down: Start node A with -noquorumlogging (available only with the debug version of the Cluster service), delete the tombstone file (quotomb.stn) from the cluster directory. You will lose all your changes subsequent to the changing of the quorum device from one drive to another. Run -fixquorum, then define a new Quorum resource.