Overview

The highest level of data protection comes from multi site replication, where a source clusters data is replicated to clusters at 2 different sites. Typically, 2 clusters are in metro location and the 3rd cluster is outside a power grid failure zone.

The goal of this failover design is to provide a choice of sites to failover and have fully automated failover between sites. In addition, failback and failover again to the same or different site will be possible, but may require some manual steps to avoid SyncIQ policies that will block the failover.

In this configuration Eyeglass can be used with Access Zone failover solution to protect the data in an Access Zone and allow a failover choice of target cluster 1 or target cluster 2 (at the 3rd site). This solution can operate at the Access Zone level and allows one or more Access Zones to be 3 site protected and other Access Zones only 2 site protected.

Video How To - Overview Multi site Access Zone Failover

Overview of Multi Site DR and Continuous Availability

This solution will support Access Zone fully automated and DFS mode multi site failover and failback. The document covers Access Zone and DFS mode in separate sections.

Overview of Multi Site Access Zone DR and Continuous Availability

This solution allows one or more independent Access Zones to have 2 possible replication targets and sites to failover. This offers maximum data protection and full automation for DNS mount path failover for SMB and NFS.

Pre-requisites

DNS also requires triple Name Server Delegation records. This is similar to Dual delegation where the SmartConnect Zones involved in the 3 site failover must now have 3 NS records pointing to all Subnet Service IP’s for each cluster and subnet involved in the failover
Igls-hints provisioned on all 3 clusters for each Access Zone must be configured
Recommended to apply on all target clusters the SmartConnect zone placeholder using igls-original-<SmartConnect Zone name on source cluster>
Replicate the same data to site B and C to keep it simple using overlapping source path policies that replicate to cluster B and C with A being the source cluster
Ensure IGLS-hints are globally unique example this avoids SPN collection issues on AD machine accounts but allows pools to be matched on the igls-xxxx-

A cluster - igls-poolx-Cluster-A
B cluster - igls-poolx-Cluster-B
C cluster - igls-poolx-Cluster-C

Use the same AD provider in all Access Zones
Ensure all clusters are in the same AD Forest
NOTE When following AD Delegation you MUST extend the delegation to the 3rd cluster and make sure each cluster is given Write Service principal Name permissions to its own Computer Account AND the other 2 cluster machine accounts.

Supported Access Zone Failover Operations

A to B or A to C

If B is Active Cluster:

A to C Access Zone failover not supported
A to C per SyncIQ Policy failover supported (manual steps for networking)

If C is Active Cluster:

A to B Access Zone failover not supported
A to B per SyncIQ Policy failover supported (manual steps for networking)