The procedure How to Configure a Failover Application Using the ScalMountPoint Resource in Oracle Solaris Cluster 4.3 Data Services Planning and Administration Guide is missing the following step, to optionally perform after Step 1:
2. If the application binaries use a NAS NFS file system and you want the resource to automatically fail over if a storage fault is detected, set the RebootOnFailure property to True.
Setting this property prevents the resource from entering a STOP_FAILED state if storage connectivity is faulted. Instead, the node where the ScalMountPoint resource resides is rebooted, and the resource restarts on another cluster node.
# clresource set -p RebootOnFailure=True scalable-mount-point-resource
The procedure How to Configure a Scalable Application Using the ScalMountPoint Resource in Oracle Solaris Cluster 4.3 Data Services Planning and Administration Guide is missing the following step, to optionally perform after Step 1:
2. If the application binaries use a NAS NFS file system and you want the node where the resource was online to abort if a storage fault is detected, set the RebootOnFailure property to True.
Setting this property prevents the resource from entering a STOP_FAILED state if storage connectivity is faulted. This STOP_FAILED state might leave application instances in a hanging state on the node that has the fault and possibly also on other nodes if the instances communicate among themselves. But if the RebootOnFailure property is set to True, the node where the ScalMountPoint resource resides would instead abort. Manual intervention is then necessary to repair the storage connection for the faulted node and restart the storage resource on that node, which then restarts the application instances on that node.
# clresource set -p RebootOnFailure=True scalable-mount-point-resource