Microsoft StorSimple – Automatic Disaster Recovery

So far our author Henrik Motzkus has concentrated his detailed look into Microsoft StorSimple here on our blog on the fundamental features. Today he will expand on a few individual aspects. One of the most important of them is disaster recovery. After basic installation of Microsoft StorSimple, the solution needs to be configured to suit the availability requirements. Henrik Motzkus takes a file server as an example to describe the technical ins and outs of this scenario.

StorSimple is a hybrid storage solution by Microsoft, designed to provide local users with simple, quick and extremely cost-efficient storage from the Microsoft Azure data center. It’s worth taking a look at  Microsoft Azure StorSimple: Easy Entry to the Hybrid Cloud to learn the basics and to acquire an initial overview of how the software works.

Architecture for automatic disaster recovery

What does the architecture look like in this scenario? The file server runs on a virtual machine (VM) with Windows Server 2012 R2. It might have a Failover Cluster Configuration. The server uses StorSimple to obtain storage as Cluster Shared Volumes (CSV), which come as iSCSI volumes. StorSimple mirrors the volumes to Azure (Cloud Snapshots), and domain users then have access. This architecture delivers excellent availability. All components are redundant. Workload and running costs are restricted to a minimum.

StorSimple and Azure Site Recovery

StorSimple and Azure Site Recovery | Source: Microsoft 

What does automatic disaster recovery need?

What do I need to know in order to make an environment fit for ‘Enterprise’ disaster recovery? This section will address the individual components that are needed in order to set up the DR. They are:

  • Cloud Appliance for iSCSI volumes in Microsoft Azure
  • Active Directory in Azure or local
  • Azure Site Recovery Vault
  • Azure Automation

In my previous blog articles I described which features Microsoft StorSimple provides. The Cloud Snapshots deserve a mention in this context. They are used to copy the entire dataset safely and in an encrypted form from a certain  iSCSI volume in Microsoft Azure Cloud to  Blob storage. Fantastic! So at least the data is secure. Now what is missing is the option of accessing the data that needs to be made available from the iSCSI volumes. This is where the  Microsoft StorSimple Cloud Appliance comes in. The Microsoft StorSimple Cloud Appliance is the virtual appliance that operates as a Microsoft Azure VM. The Cloud Appliance provides the same features as the Hardware Appliance, and makes iSCSI volumes available to the servers. But the following consideration is imperative: a Cloud Appliance can manage no more than 64 TB of data. So the iSCI volumes need to be spread over several Cloud Appliances if the Hardware Appliance holds more than 64 TB of data. The reason for this is that the Cloud Appliance is a standard VM in Azure. And as we know a VM supports a maximum of 64 hard disks with 1023 GB each.

Step-for-step guide to automatic disaster recovery

1. Microsoft Azure Active Directory

Now there needs to be a way to authenticate users. There are the following options, depending on the complexity of the environment:

  • The Domain Controller can take  Azure Site Recovery to mirror the entire VM for less complex environments in which only one Domain Controller is running and there is a small number of users. The mirror is then simply booted in the event of a disaster. Afterwards all of the changes in the structure in Azure need to be repeated in the company’s data center. This way of doing things does not particularly appeal to me and in some cases it will not even work.
  • It is advisable to set up another DC in Azure to accommodate large environments with many DCs (in a forest), users and a high rate of changes in the AD. This DC then becomes a part of the overall structure, automatically mirroring all changes in the structure back to Microsoft Azure as part of the  Replication Topology.

2. Microsoft Azure Site Recovery

Then it is necessary to make sure that the file servers on which the users access the shares (by DFS, Failover Cluster, etc.) are also available during the disaster. In this case, Microsoft Azure Site Recovery is used to replicate the corresponding VMs in the Azure Data Center as a precise copy down to the last bit. This way they can also be booted in the event of a disaster. The following diagram outlines basically how Microsoft Azure Site Recovery works in this specific scenario.

How Azure Site Recovery works

How Microsoft Azure Site Recovery works | Source: Microsoft

This needs the Azure Site Recovery Provider that coordinates disaster operations on the Hyper-V host. It uses port 443 to communicate with the Microsoft Azure Data Center. Data protection people will be delighted. A Recovery Services Agent on each VM is also required. Where needed, it runs scripts on the actual VMs in order to execute the operations during a disaster.

3. Microsoft Azure Automation

Last but least there will have to be the opportunity for the automatic execution of a failover process. Microsoft Azure Automation is predestined for this task. Users can set up an Azure Automation Account or select a ready-made runbook from a pool. After adding the runbook to the personal account it only takes a mouse click to launch several operations.

The test failover does not impact the local servers at all. The VMs are merely booted in Azure and the volumes are mapped to the VMs. Before that happens the StorSimple volumes are cloned and attached to the Virtual Appliance, thus producing the perfect playing field.

The planned failover tries the ‘friendly approach’ to shutting down the VMs in the local data center in order to then boot the resources in Azure. To do this an existing Cloud Snapshot is mapped to the Virtual Appliance. Voilà!

The unplanned failover then simply boots the VMs in Azure and executes a failover of the StorSimple volume container. Afterwards the volumes are available to the VMs, and access to the data is restored for the file server.

Automated Disaster Recovery Solution using Azure Site Recovery

Automated Disaster Recovery Solution using Azure Site Recovery | Source:  Microsoft

Ready-made runbooks in particular make the whole process very simple. The sample scripts are written and can be modified to suit the specific scenario.

Summary and outlook

I analyzed the numerous benefits of the Microsoft StorSimple solution in earlier blog articles. The hybrid solution presented here promises excellent availability for the file services provided, extremely simple administration and good potential to cut costs. After all, the Microsoft Azure data center is used as a disaster location. So a second, dedicated location running failover hard and software is therefore obsolete.

In future, Microsoft will place a strong focus on developing the hybrid data center solution. In Azure Site Recovery, Azure StorSimple or Azure Active Directory there are already finished solutions for ideal connections between the local infrastructure and Azure.

About the author

Henrik Motzkus is IT Architect for Enterprise Datacenter Infrastructure, Project Manager for Infrastructure Projects and Expert for IT Operations at COMPAREX. Here, he focuses on the vendors EMC, Cisco, VCE and Microsoft, catering to both cloud and on-premise solutions. He uses his 20 years of experience in datacenter solutions to collaborate with customers in creating integral solutions that draw potentially on every area of available technologies.

 LinkedIn |  Xing |  Blog (German)

Henrik Motzkus, IT Architect at COMPAREX

Leipzig, 25.08.2016

Stay Up-to-date

For regular updates and articles from COMPAREX, click below to follow us:

 Follow us on LinkedIn

Reach Henrik through

 LinkedIn
 Xing
 Blog

Share this Article

Leave a Comment

Do you have a question or remark on this article you want to share with us?
 Post it here.