Disaster Recovery and Automated Protection in the Cloud with Azure Site Recovery

Business Case Story

The client used a secondary datacenter for Disaster Recovery of their primary datacenter. They needed a solution in Azure which to replace their secondary datacenter.

Solution Story

To implement an Azure Site Recovery solution for Disaster Recovery of the primary datacenter servers and to migrate Disaster Recovery infrastructure from the client’s secondary datacenter environment, we delivered a solution with a CSP Subscription in Azure. It was provisioned by Zelusit’s managed service team.

Figure 1: Data Recovery Infrastructure with Azure

Our team used the existing express route between EA Subscription and On-premise to establish connectivity. In the CSP Subscription, one resource group was created to hold all required resources: a recovery vault, storage accounts, virtual networks. Replication policies were also created during the setup. All VNets were connected via virtual network peering. The connectivity between the two Azure subscriptions was established via cross-subscription VNet peering, which also made possible the connectivity between on-prem and CSP subscription.

The on-prem Domain Controller was replicated in EA Subscription as a Backup Domain Controller (BDC). BDC was added as a DNS server in all CSP VNets created for this project. In case of disaster recovery, the virtual machines created automatically after failover in CSP subscription, are going to use it.

All necessary firewall rules were added to the on-prem firewall and EA Subscription’s firewall client to allow the communication related to the implementation of this project from on-prem to Azure and vice versa.

Part of the preparation was to configure an on-prem configuration server which to serve as Configuration (CS), Process (PS) and Master Target server (MTS) for all in scope servers in the Datacenter.

To perform failback operation, a separate Process Server in CSP Subscription was configured.

Recovery plans based on server’s type were created after all in scope servers were replicated.

Challenges

There were a couple of challenges during the setup of the on-premise configuration server and the mobility service installation, also part of the preparation for initial replication of the servers.

All tasks related to the on-prem configuration and setup of firewall rules was done by the Cloud Architect from customer’s side with the help of the Azure Lead and the Azure Consultants. This was required from the client.

Zelusit followed Microsoft documentation to setup and perform the ASR implementation. Our consultants performed the following high-level steps to cover the full cloud migration cycle:

• Replicate and protect on-prem server (on-prem to Azure direction) – replicates on-prem VM/physical server to Azure.
• Failover (on-prem to Azure direction) – creates and brings up an Azure VM using a selected recovery point.
• Commit – completes the failover operation and deletes all available recovery points for the failed over server.
• Re-protect (Azure to on-prem direction) – replicates Azure VM to on-prem. Physical servers replicated to Azure using Site Recovery can only fail back as VMware VMs.
• Failover (Azure to on-prem direction) – fails back from Azure to on-prem, shuts down Azure VM.
• Commit – completes the failover operation and deletes Azure VM of the failed over server.
• Re-protect (on-prem to Azure direction) – replicates on-prem VM to Azure.

Zelusit performed a successful test failover for each server in scope and actual failover/failback for 2 servers which completed customer’s requirement.

Benefits Story

The solution proved that the disaster recovery in the cloud can eliminate the secondary datacenter as well as all accompanying costs and maintenance complexity. The solution enabled our client to perform seamless disaster recovery drills, without affecting production environments. It serves as a single-point-of-management for replication, failovers, and failbacks. Our team provided detailed step-by-step technical documentation.

• Customized failovers and fast recovery of multi-tier applications running on multiple VMs via Recovery plans.
• Continuous health monitoring and failover status.
• Protection for the virtual machines and the physical server.
• Recovery time objectives (RTO) and recovery point objectives (RPO – 60 min) within organizational limits.
• Data reliability by using recovery points with application-consistent snapshots.