Failing over is useful as a test procedure or in order to minimize disruption while production storage is under maintenance. This procedure does not apply to GAD. GAD failover is invisible to the NAS server.
CAUTION:
Failover always involves a brief loss of service.
- Unmount all file systems on the span you want to fail over.
- Use your storage configurator to ensure that the S-Vols contain an up-to-date copy of the data on the P-Vols. If relationships are synchronous, this means ensuring that the relationships are healthy; otherwise, it can mean running an update command or script.
- Use your storage configurator to promote the current S-Vols to the primary role and demote the current P-Vols to the secondary role.
- Run the pn all span-list command until it shows that the span is healthy again throughout the cluster.
- Remount the filesystems. To fail back, repeat the procedure.