Two-node cluster with tape devices on a shared SCSI bus

In this configuration, cluster-aware Backup Exec is installed on the controlling node, failover occurs to designated nodes in the cluster, and tape devices are attached to a shared SCSI bus that is separate from any shared SCSI bus for disks.

Figure: Two-node Cluster with Tape Devices on a Shared SCSI Bus

Two-node Cluster with Tape Devices on a Shared SCSI Bus

Because each node creates a unique tape device name for the same device, if the drive is not serialized, this configuration requires you to create a device pool that includes the tape device name used by each node in order for jobs to run when failover occurs.

When failover occurs, a SCSI bus reset is issued. Therefore, tape devices and shared drives should not be connected to the same SCSI bus; each should be connected to separate SCSI buses.

Note:

If you are using a serialized tape device in a shared SCSI cluster configuration, media that is orphaned in a device because of a failover will be ejected from the tape device. If you are using a tape device that is not serialized, you need to manually eject the media from the device or reboot the device

More Information

Creating device pools for Microsoft Cluster Servers

Configuring a shared SCSI bus for tape devices

Multi-node clusters on a fibre channel SAN with the SAN SSO