You can configure the cluster to automatically recover from an abrupt total cluster failure. The automatic cluster takeover is triggered if an instance attempts to connect to the cluster and:
You have enabled automatic cluster takeover.
The quorum device indicates the cluster is running, but the instance detects no heartbeat.
The automatic cluster takeover configuration parameter enables the instance attempting to connect to the cluster to form a new cluster, starting the cluster coordinator and recovering the databases. The syntax for automatic cluster takeover is:
sp_configure "automatic cluster takeover", [1 | 0]
Setting automatic cluster takeover to 1 enables the automatic cluster takeover (the default). Setting it to 0 disables automatic cluster takeover.
See “Starting the cluster” for information about manually restarting the cluster.
In environments that have i/o fencing enabled, automatic cluster takeover is guaranteed to be a safe operation. In environments that don't have i/o fencing, a malfunction of the algorithm could introduce data corruption. Therefore this configuration parameter is present to disable the algorithm. However there is always a risk of data corruption in environments without i/o fencing, and disabling automatic cluster takeover does not mitigate all of those risks.
