The following best practices are recommended:

• Keep source volumes and snapped volumes from sharing the same physical disks. If this is not maintained, then any attempt to split the snapshot volume from the original volume fails.

• Most hardware and software providers have some limitation about the types of volumes that are transportable. Therefore, Symantec recommends that you use offhost backup jobs only for backing up data for which all dependent volumes, or mounted volumes, can be imported and deported.

• Using offhost backup to back up Veritas Storage Foundation for Windows (VSFW) volumes requires that snapshot volumes in shared storage be transferred from host to host. Make sure that VSFW volumes that are backed up with offhost backup reside in VSFW disk groups that have either the "private protection" or "cluster disk group" disk group property. Private dynamic disk group protection and cluster disk group property settings use hardware locking techniques to protect a dynamic disk group that is located on shared storage from being accessed by other hosts that are connected to the shared storage pool.

• The offhost backup will fail if any one volume that you select for backup is only supported by a Microsoft Volume Shadow Copy Services (VSS) provider and cannot be imported or deported, or if the required VSS hardware provider is not on a Symantec-approved compatibility list. You can choose to continue the backup if the offhost backup fails.

  You can find a list of compatible devices at the following URL:

  http://entsupport.symantec.com/umi/V-269-2

• The Hitachi Raid Manager log cannot be on a volume being snapped. Hitachi executes I/O to its Raid Manager log file during the snapshot commit process, and the VSS coordinator blocks I/O to any drive being snapped. Therefore, if the log directory for Raid Manager is on the volume that is being snapped, then log I/O is blocked and the snap process is deadlocked.

• If the Central Admin Server Option (CASO) is installed, for jobs that use offhost backup, you must manually select the destination device that will run the job rather than allowing the job to be delegated by the central administration server. Otherwise, the job could be delegated to a media server that does not have offhost capability.

  See Grouping managed media servers in pools in CASO .

• When performing an offhost backup using a VSS hardware provider in a Microsoft Cluster Server (MSCS) or Veritas Cluster Server environment, the media server and the remote computer must not be in the same cluster group. The cluster applications cannot support devices' logical unit numbers (LUNs) that have duplicate signatures and partition layouts, therefore, the snapshots containing the LUNs must be transported to a host, or remote computer, that is outside the cluster.