Skip to main content

Ways to implement SVM disaster recovery in SAN environments

The storage virtual machine (SVM) disaster recovery provides support for data recovery at the SVM level if an SVM becomes inaccessible.

The primary SVM is the SVM requiring disaster recovery support. The secondary SVM is an SVM in another cluster that is peered with the primary SVM to provide disaster recovery support. The cluster containing the secondary SVM does not have to have the same number of nodes or the same number of FC and Ethernet ports as the primary SVM.

If the -vserver-dr-protection option of the volume is set to unprotected , the SVM disaster recovery does not replicate this volume at the destination SVM. Existing volumes and newly created volumes on the source SVM are protected by default.

Volume, configuration, and metadata is replicated to the secondary SVM at regular scheduled intervals. If the primary SVM becomes inaccessible, the secondary SVM can be brought online. Persistent reservations are not copied to the secondary SVM. This means that certain hosts must either be rebooted or have persistent reservations reset after the secondary SVM becomes active. The following hosts require a reboot or a persistent reservation reset:

  • Veritas

SVM disaster recovery can be implemented in two ways.

Identity preserve SVM disaster recovery​

With identity preserve SVM disaster recovery, volumes, LIFs, and LUNs in the secondary SVM are not visible to hosts until the entire disaster recovery operation has been successfully completed.

The volumes, LIFs, and LUNs in the secondary SVM have the same identity as the corresponding volumes, LIFs, and LUNs in the primary SVM. Therefore, the primary SVM and secondary SVM cannot be visible to the host simultaneously.

Identity preserve SVM disaster recovery does not require FC port zoning.

Identity discard SVM disaster recovery​

With identity discard SVM disaster recovery, the volumes, LIFs, and LUNs in the primary SVM remain visible to the host in read-only mode for the duration of the disaster recovery event.

The volumes, LIFs, and LUNs in the secondary SVM do not have the same identity as the corresponding volumes, LIFs, and LUNs in the primary SVM. Therefore, the primary SVM and secondary SVM can be visible to the host simultaneously.

To use identity discard SVM disaster recovery, the primary SVM FC LIFs must use WWPN zoning.