Designlet: Site Recovery Manager for Azure VMware Solution

Introduction

Azure VMware Solution (AVS) combines VMware compute, networking, and storage running on top of dedicated bare-metal hosts in Microsoft Azure. VMware Site Recovery Manager (SRM) for Azure VMware Solution is an add-on that customers can purchase to protect their virtual machines in case of disaster. SRM for AVS allows customers to automate and orchestrate the failover and failback of virtual machines between an on-premises environment and AVS, or between two AVS sites.

Summary and Considerations

Use Case	Site Recovery Manager protects organizations from disaster by allowing customers to replicate virtual machines, create dynamic recovery plans, perform non-disruptive failover tests, and recover virtual machines at a secondary site.
Pre-requisites	ExpressRoute connectivity between on-premises and Azure VMware Solution ExpressRoute Global Reach between source and target Azure VMware Solution private clouds DNS resolution between on-premises SRM and virtual cloud appliances Site Recovery Manager 8.3.x
General Considerations/Recommendations	Topologies supported: On-premises to AVS Primary AVS site to Secondary AVS site One-to-Many and Many-to-One are not supported Custom SRM Extension ID is not supported Array-based replication, VVOLs, and storage policy Protection Groups are not supported IP Customization using SRM command-line tools is not supported If IP Customization Rules have been defined for network mappings between the AVS environment and the on-premises environment, these rules will not be applied on failback from the AVS environment to the on-premises environment due to a known issue with SRM 8.3.0. You can work around this limitation by removing protection from all VMs in the Protection Group and then reconfiguring protection on them prior to initiating the failback. Scale limitations per private cloud: Number of protected VMs: 1,000 Number of VMs per recovery plan: 1,000 Number of protection groups per recovery plan: 250 RPO values: 5 minutes or higher Total number of VMs per protection group: 500 Total number of recovery plans: 250
Performance Considerations	Configure the RPO for each virtual machine appropriately based on business need and data change rate Ensure sufficient bandwidth is available to support desired replication Calculate bandwidth for replication: Identifying the average data change rate within the RPO, the average change rate over a longer period, and then dividing by the RPO. Calculate how much traffic the data change rate generates in each RPO period. Measure the traffic against your link speed.
Network Considerations/Recommendations	Ensure appropriate network ports are open for SRM and vSphere Replication.
Cost Implications	The `Site Recovery Manager for Hyperscalers` SKU is sold in packs of 25 VMs for 1- and 3-year terms by VMware and VMware partners. Existing on-premises licenses cannot be reused. Egress charges may apply to traffic communicating from Azure VMware Solution to an on-premises environment. Cross-regional and cross-zonal traffic in Azure may incur additional costs.
Document Reference	SRM on AVS Microsoft Documentation VMware SRM Documentation VMware vSphere Replication Documentation
Last Updated	November 2021

Implementation

The following instructions are for use with an on-premises to AVS deployment scenario.

Deploy Site Recovery Manager

Follow the documentation to deploy VMware Site Recovery Manager and vSphere Replication on-premises.
In your Azure VMware Solution private cloud, under Manage, select Add-ons > Disaster recovery
From the Disaster Recovery Solution drop-down, select VMware Site Recovery Manager (SRM) – vSphere Replication
Provide the license key, select agree with terms and conditions, and then select Install
- You can use a production license key or leave this field blank to use evaluation mode. The evaluation mode is valid for 60 days. Once a production license is obtained, you’ll be able to update the license key in the AVS portal.

Once the deployment of the SRM appliance is complete, you can deploy the vSphere Replication appliance(s).

Deploy vSphere Replication

Each replication appliance can accommodate up to 200 protected virtual machines. You may scale in or scale out the number of appliances to meet your needs.

From the Replication using drop-down, on the Disaster recovery tab, select vSphere Replication
Move the vSphere server slider to indicate the number of replication servers you want based on the number of VMs to be protected, then select Install

Configure Site Pairing

In the on-premises vSphere Web Client, click Site Recovery > Open Site Recovery
Click the New Site Pair button
Select the first site from the list. Enter the address of the Platform Services Controller for the Site Recovery Manager Server in AVS, use the cloudadmin@vsphere.local credentials when prompted, and click Next
Select the vCenter Server and the services you want to pair and click Next
On the Ready to complete page, review the pairing settings, and click Finish

Additional Resources

Authors and Contributors

Jeremiah Megie, Principal Cloud Solutions Architect, Cloud Infrastructure Business Group, VMware
Steve Pantol, Senior Technical Marketing Architect, Cloud Infrastructure Business Group, VMware