Every company has the potential to be hit with hardware failure and system downtime. If disaster strikes your organization,...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
how quickly could you restore service -- even in a degraded state?
The Azure Site Recovery service uses Microsoft's cloud platform to prevent a halt in operations when issues arise. Azure Site Recovery moves workloads to and from different data centers -- as well as both public and private clouds -- to keep key services online and available.
What is Azure Site Recovery?
The Azure Site Recovery service has two elements:
- The software and connections move VMs and services between two private data centers -- either owned or rented by your organization -- including Hyper-V and VMware VMs.
- The Azure public cloud service acts as a data center stand-in and provides hot site disaster recovery capabilities. The Azure Site Recovery service also supports the hypervisors on Hyper-V and VMware vSphere. Azure Site Recovery does not work with the Xen hypervisor.
New Azure portal offers advanced management
At one time, administrators needed PowerShell to set up Azure Site Recovery to use Azure Resource Manager style deployments. IT shops can now use the new Azure portal to set up a new Azure Site Recovery environment, including a recovery vault.
This update enables IT to specify different VM sizes within the same account and set up fine-grained access to each resource based on user roles. Only the new portal supports fresh deployments, but it also can manage and support any existing deployments that began via the "classic" portal.
How to set up Azure Site Recovery
In addition to an Azure subscription, the organization needs an Azure storage account that holds data replicated from on-premises servers.
Log into the new portal to create a Recovery Services vault inside the storage account. Select New > More Services > Monitoring + Management > Backup and Site Recovery (OMS) to create VMs with replicated data; these failed-over Azure VMs also need access to an Azure network.
The cost of peace of mind
Microsoft bills Azure Site Recovery based on each protected instance. This covers the software and orchestration, as well as the use of Azure compute if you run protected virtual machines within Azure as your hot spare location, storage for replicated data and outbound data transfer in failback scenarios.
If you recover to your own data centers, the price per virtual machine is $16 a month. If you recover to Azure, it's $25 a month. Microsoft provides a 31-day free trial.
VMware shops will need a local VM to run the configuration server role that coordinates the data and communication with Azure and also handles the data replication processes. This VM is the process server and functions as a replication gateway -- it caches, compresses and encrypts replication data, then sends it to Azure. The process server discovers other VMs and adds them to a replication configuration. The configuration server also acts as the master target server, which handles the replication after a disaster concludes and roles shift from Azure back to the on-premises locations.
Windows and Hyper-V shops need either System Center Virtual Machine Manager in the on-premises environment to manage the VMs or the Site Recovery Provider that communicates with the service over the internet. They also must install the Recovery Services agent on non-Virtual Machine Manager hosts to manage data replication.
How does it work?
From there, the Azure Site Recovery service does most of the grunt work. It manages replication based on pre-programmed cycles of 1 minute, 2 minutes, 15 minutes and so on. After the initial seeding, Azure Site Recovery performs delta replication to save bandwidth. You can set up "exclude disks" to avoid replication of temporary files and page files.
Remember to set up a recovery plan that instructs the services where VMs go, on what schedule and in what order; this creates a recipe to follow if a disaster or business interruption occurs. You can then trigger a failback once the interruption concludes and return services to their normal operation and location.
When to consider Azure Site Recovery for cloud disaster recovery
Microsoft adds tagging tools to Azure Site Recovery
Automating data replication with Azure Site Recovery