Clustered VTSS from StorageTek provides high data availability to users of the VSM system. A clustered VTSS allows customers to replicate virtual tape volumes (VTVs) from one VTSS to another. The two VTSSs operate together to provide improved data availability without additional host or operator involvement.
|ACS||Automated Cartridge Store|
|RTD||"Real" Tape Drives|
|VSM||Virtual Storage Manager|
|VTV||Virtual Tape Volume|
|VTSS||Virtual Tape Sub System|
The two VTSSs in a cluster can be attached to a single ACS at one location or two ACSs in one or more locations. The VTV replication is started automatically at end-of-volume. Clustered VTSS can be utilized to provide automatic copying of VTVs to a remote location automatically.
Originally, when two VTSSs were connected to become a cluster, one VTSS was configured as the primary system and the other as the secondary system. Under normal conditions, the primary VTSS is responsible for supporting the production tape workload. The primary VTSS also replicates the VTVs across a channel connection between the primary and secondary VTSSs.
Migration is the process of copying data from the disk buffer to the physical tape media. Migration can occur from either the primary or secondary VTSS, but is typically performed by the secondary which frees up resources on the primary. As soon as VTVs are replicated to the secondary VTSS, they are queued for immediate migration.
If the primary VTSS becomes unavailable, it can be varied offline. This allows the workload to continue using only the secondary VTSS. Once the primary is available, it can be varied online. This will return the workload to the primary VTSS. When the primary is restored back online, VTCS will automatically re-synchronize the contents of the two disk buffers.
Current versions of the Sun/STK software provides bidirectional cluster support. Just as in the initial primary/secondary configuration, two peer VTSSs are connected by one or more links. In a bidirectional cluster, the system administrator can allow replication in either direction from VTSS to VTSS across a cluster link. In this configuration, each VTSS services a subset of the production workload and functions as a “warm standby” for the other.
This document was printed from http://recoveryspecialties.com/