When you have the need to implement Disaster Recovery scenarios towards Azure particularly in complex environments, through the solution Azure Site Recovery (ASR), you can use the Azure Site Recovery Deployment Planner, recently released by Microsoft, to make a detailed assessment of the on-premises environment. The tool is designed to cover both Hyper-V and VMware environments . In this article, we will detail the use of the tool when you are trying to activate a Disaster Recovery plan with replication of VMware virtual machines to Azure.

What is the use of this tool?

ASR Deployment Planner performs a detailed assessment of the on-premises environment, aimed at using the solution Azure Site Recovery (ASR), and provides elements to consider in order to contemplate the various operations needed to effectively implement the plan of DR: replica, virtual machine failover and DR-Drill. The tool also performs an estimate of Azure resources required for the protection of on-premises virtual machines, reporting information about costs for the use of ASR.

In the presence of VMware environments if you have the need to address real migration scenarios towards Azure, the most appropriate tool to use to carry out the assessment of the environment is Azure Migrate.

How to use the tool?

The use of ASR Deployment Planner involves two main stages. The first of profiling, during which the necessary information is collected from the environment VMware, and the second of report generation to perform the analysis.

ASR Deployment Planner can be downloaded at this link. This is a compressed folder whose contents should be copied on the system on which you intend to run the tool. ASRDeploymentPlanner.exe is the command line tool that must be executed with the appropriate parameters, there is no required installation.

Profiling and measurement of throughput

The machine on which you intend to make the profiling or calculating the throughput must meet the following requirements:

• Operating System: Windows Server 2016 or Windows Server 2012 R2.
• Hardware requirements: 8 vCPUs, 16 GB RAM and 300 GB HDD.
• Software Requirements: .Net Framework 4.5, VMware vSphere PowerCLI 6.0 R3, Visual C++ Redistributable for Visual Studio 2012.
• Internet access to Azure.

In addition the following conditions are required:

• Presence of an Azure storage account (only if you want to calculate the throughput).
• VMware vCenter statistics level set at level 2 or higher.
• Ability to connect to vCenter server/ESXi host on port 443.
• User with at least Read-only permission to access the VMware vCenter server/VMware vSphere ESXi.

In general it is a good idea to perform the profiling and the calculation of throughput on the Configuration Server you intend to use, or on a system with similar characteristics.

The tool is able to perform the profiling only for virtual machines with RDM and VMDK disks. There is no collection of information of VMs with iSCSI or NFS disks; in this regard it should be noted that Azure Site Recovery does not support virtual machines with these types of disks in a VMware environment.

During the profiling activity the tool connects to the vCenter Server or vSphere ESXi host to collect performance data for virtual machines. This implies that the data collection activities has no impact on the performance of virtual machines because there is no direct connection. The profiling is done once every 15 minutes as not to impact on VMware systems, but the query that is performed, however, collects performance data for all the time interval.

The profiling activity requires the presence of a text file containing the list of virtual machines (a name or an IP address for each row) you intend to examine. This file you can create it manually or, with the following commands, performed from the VMware vSphere PowerCLI console, you can extrapolate the list of all virtual machines on the vCenter or on vSphere ESXi host.

Figure 1 - Extrapolation of VMs from the vCenter

Figure 2 – Example of the file containing the list of VMs

At this point you can begin the profiling process. For production environments it is recommended to run it for at least a week, so you have a sufficiently long period of observation to achieve accurate profiling. To get the complete list of required and optional parameters, you can run the following command: ASRDeploymentPlanner.exe-Operation StartProfiling /?.

Among the optional parameters you can also specify an Azure Storage Account with its key to calculate the throughput that Site Recovery can reach during the replication process to Azure.

Figure 3 -Example of running the profiling

If the server, on which profiling process starts, be rebooted or goes in crash, the data collected would remain and you can simply restart the tool.

The tool can also be used for throughput calculation.

Figure 4 - Example of throughput measurement

The process of measuring the throughput will upload files with extension .VHD on the storage account specified. Upon completion of the upload these files are removed automatically from the storage account.

Report Generation

The machine on which you want to generate the report must have installed Excel 2013 or a higher version.

After the profiling process you can generate the report containing the output of the assessment. When you create the report, you must run the tool in report-generation mode. In this case to consult all the possible parameters you should run the command ASRDeploymentPlanner.exe -Operation GenerateReport /?.

Figure 5 - Example of the command for the report generation

The report generated is called DeploymentPlannerReport_xxx.xlsm within which you can see different information, including:

• An estimate of the network bandwidth required for the initial replication process (initial replication) and for delta replication.
• The type of Storage (standard or premium) required for each VM.
• The total number of storage accounts (standard and premium) required.
• The number of Configuration Server and Process Server you need to implement on-premises.
• The number of VMs that can be protected in parallel to complete the initial replication at any given time.
• Estimating the throughput attainable by ASR (on-premises to Azure).
• An assessment of the supported virtual machines, providing details about the disks (number, its size and IOPS) and the type of the OS.
• Estimation of DR costs, for use it in a specific region of Azure.

Figure 6 - Home page of the generated report

To obtain detailed information concerning the analysis of the report please visit the Microsoft's official documentation.

In addition to being present in the home page of the report a summary of the estimated costs, there is also a specific tab containing the details of the cost analysis.

Figure 7 – Section on cost estimates in the report generated

For more details on the information and its interpretation, you can check the official documentation.

Conclusions

Azure Site Recovery Deployment Planner is a very useful tool that, making a detailed assessment of the on-premises environment, allows not to omit any aspect to achieve in the best way a Disaster Recovery plan towards Azure, using Azure Site Recovery (ASR). This tool also allows you to have with great precision an estimate of the costs that you need support for the disaster recovery plan, so you can make the necessary evaluations.

In azure, there is the possibility of using Azure Site Recovery (ASR) to implement easily an efficient disaster recovery strategy by enabling replication of virtual machines among different regions of Azure. Although in Azure are present integrated mechanisms to deal with localized hardware failures, it may be appropriate to implement a solution that can ensure applications compliance , performed on virtual machines in Azure, against both catastrophic events, such as earthquakes or hurricanes, that software issues that may impact on the functioning of an entire region of Azure. This article will show you how to configure a virtual machine replication and how to enable a disaster recovery scenario.

This feature has been defined one-click replication because of its simplicity, it is currently in public preview and it is usable in all the Azure regions where ASR is available.

Before you enable this functionality is essential to ensure that the necessary requirements are met and to do that you can see the compatibility matrix for the replication scenario of virtual machines among different regions.

By accessing the Azure Portal it is possible to select the virtual machine that you intend to replicate and perform the configuration in the section Disaster recovery:

Figure 1 – Disaster Recovery Section of the VM

Selecting Disaster Recovery shows the following configuration panel:

Figure 2 – VM replication configuration panel

The first required parameter is the target region where you want to replicate the virtual machine. The replication activation process also create the necessary Azure artifacts (Resource Group, Availability Set if used by the selected VM, Virtual Network and Storage accounts) or you can select them at will if they were created earlier.

Figure 3 – The resources needed in the region target

The replication process also requires the presence of a Cache Storage Accounts in the source region that is used as a temporary repository to store changes before they are reported in the storage account defined in the target region. This is done to minimize the impact on production applications that reside on the replicated VM.

Figure 4 - Cache Storage Account in the replication process

Always in the configuration panel is required which is proposed Vault Recovery Services use creating a replication policy that defines the recovery point retention and the rate at which consistent snapshots are made at the application level.

By selecting Enable Replication will begin the creation process of Azure resources required, the VM is registered in the selected Recovery Services Vault and replication process is activated.

The Disaster Recovery section lists details about the replication and it is possible to perform a failover or a test failover:

Figure 5 - Details relating to the replication process of the VM and activation of the failover process

The procedure Test Failover Specifies which recovery point using between: latest, latest processed, latest app-consistent or custom. In addition it is possible to select in which virtual network attest the virtual machine during the test failover in order to perform the test without generating any impact on the production systems.

Figure 6 – Test Failover of a VM

Similar the Failover panel that allows you to specify only which recovery point to use as the network on which attest the machine has already been defined in the configuration phase.

Figure 7 – Failover of a VM

Only when you start the Failover process affected virtual machines are created on the target resource group, attested to the target vNet and configured in the availability set appropriate when used.

Figure 8 – Failover process

Conclusions

Thanks to this new feature introduced in Azure Site Recovery it is possible to activate with ease replication of virtual machines in different Azure regions, without the necessity of having expensive secondary infrastructure to activate a disaster recovery plan.