September 2015 Leveraging the power of Microsoft Azure to build your (Hyper-V) datacenter Disaster Recovery Plan Understanding the possibilities of Microsoft Azure Site Recovery AUTHOR: MVP & MCT PETER DE TENDER
Contents Introduction 3 1. Disaster Recovery is a top priority 3 2. Some key definitions 4 2.1 What is a disaster? 4 2.2 What is disaster recovery? 4 2.3 RPO / RTO 4 3. Familiar DR Solutions 4 4. And then there was cloud 5 5. Azure Site Recovery features 6 6. How Azure Site Recovery Works 7 6.1 Characteristics 7 6.2 Configuration 7 7. Outcome 8 8. Monitoring 8 9. Summary 9 About Peter De Tender 10 About Savision 10
Introduction High availability and disaster recovery are key components of an Enterprise s IT environment. More and more companies are adopting Hyper-V on Server 2012 R2 for their private cloud solution. This whitepaper will explain how you can leverage the power of Microsoft Azure, to build a Hyper-V based disaster recovery datacenter. That s right! You can migrate your in-house VM s to Microsoft Azure, guaranteeing uptime and availability of the machine and the applications to your end-users. Not using Hyper-V yet? No worries, Azure Site Recovery can also act as failover for your VMware or physical host platform. Learn from Peter De Tender, independent technical speaker, valued Microsoft Infrastructure Architect, Microsoft Certified Trainer and Windows IT Pro MVP, about the core features of Azure Site Recovery, extended with insights from Savision s CTO Rob Doucette, explaining how Savision solutions can assist you in monitoring the ASR topology by showing real-time feedback. 1. Disaster Recovery is a top priority When talking about disaster recovery to any of my customers, a small SMB or a large enterprise, they all put disaster recovery as a top priority on their IT budgets. And actually, disaster recovery should not be seen as an IT Pro business enabler, but something that is of vital importance to any organization. A bit more concrete, the following workloads can be listed: - 48% of the organizations will invest in server consolidation; - 37% of the customers see VM backup/recovery as a TOP priority; - 32% of the organizations place VM replication for DR as a TOP3 feature on the list; (This information is based on a survey I did with several of my customers) When comparing these numbers to more official information from sources such as ComputerWeekly.com, and taken from a different angle where they actually compared the disaster event in relation to the impact on the businesses, the following statistics are shown: - Over 70% of businesses involved in a major disaster (e.g. Fire, earthquake, etc.) either do not reopen their business, or are closed within 18 months after the disaster occurred. - 80% of businesses suffering a datacenter disaster who don t have a disaster recovery plan go out of business; - A study from Gartner Inc. found out that 90% of companies that experience data loss go out of business within 2 years. 3
These numbers clearly show that having a disaster recovery plan is crucial for your business. 2. Some key definitions 2.1 What is a disaster? A disaster is an unexpected event with destructive consequences that requires some recovery process to be solved, having a cause that is either: - Natural - Human - Technical 2.2 What is disaster recovery? Disaster recovery is the set of processes, policies and procedures that deal with preparing for recovery or continuation of technology infrastructure critical to an organization, during the disaster or after the disaster has occurred. 2.3 RPO / RTO RPO = Recovering Point Objective refers to the amount of data loss that is tolerable. It represents points in time of the most recent backup, prior to system failure. RTO = Recovery Time Objective refers to the system downtime that is acceptable. This includes the time to detect the failure, prepare backup servers, initialize the failed application(s) and reroute requests to a backup site or restore data in full. 3. Familiar DR Solutions Disaster recovery as such is nothing new. Since the early days of mainframe computing and systems, IT admins were busy handling tapes, and making sure all data that was vital to the business got stored on a second medium, preferably outside of the main computer room. Complex and very expensive solutions were put in place to have a secondary site available, where a full or recent copy of the systems and data was available. Besides tape backups, the systems also were equipped with redundant components, actually trying to provide a disaster recovery solution. Think of the different RAID-set technologies for disk subsystems, allowing to have the system or data still available, in case of a disk (one or more) crash; or redundant power supplies, or redundant network interface cards and so on. 4
Thinking a bit more out of the server itself, data was more often stored on NAS (Network Attached Storage) or SAN (Storage Area Network) solutions. Depending on types and budget, they can provide built-in disk storage replication mechanisms between both storage solutions. This replication could be synchronous (immediate) or asynchronous (short medium delay). While all this was (and to a certain extend probably still is ) great at that time, it didn t end there. Customers not only want to have a copy of the data available, they also want full systems available at the disaster recovery site. Technical solutions like clustering could help here. Replication on hypervisor level is another popular one (think of Microsoft s own Hyper-V Replica in Windows Server 2012 R2 as a good example, which replicates virtual machines from one Hyper-V host to another, even when having low bandwidth capacity available). 4. And then there was cloud All of the above is valid when talking about local datacenters, MPLS networking connections, fibre channel interfaces between servers and storage solutions but what happens when your virtual machines are somewhere in the cloud? How do you deal with disaster recovery in that case? That s exactly where Azure Site Recovery comes in play. Starting from a pure cloud perspective, adopting public cloud as a mechanism for disaster recovery can be motivated using the following: - Commodity cloud solutions are available in the market - It s using cloud payment mechanism OPEX instead of CAPEX - Cloud storage is (in most cases) well affordable - It s cloud technology, but based on proven technology 5
5. Azure Site Recovery features Azure Site Recovery (ASR) saw the light about 2 years ago, when it was still called Azure Hyper-V Recovery manager. The key functionality at that time was using Azure Hyper-V Manager as an orchestrator to initiate and control failover between two Hyper-V datacenters, the current version of ASR provides the following disaster recovery scenarios: Summarized overview of supported replication Replicate To Replicate From Description Azure Hyper-V Site VMM Server Physical Server VMware virtual machine Replication between one or more Hyper-V hosts on-premises and Azure; no VMM server needed. Replication between one or more Hyper-V hosts on-premises in a VMM cloud and Azure. Replicate a physical Windows or Linux server to Azure Replicate VMware virtual machines to Azure Summarized overview of supported replication Replicate To Replicate From Description Secondary Datacenter VMM Server (Cloud) VMM Server with SAN Single VMM Server Replication between one or more Hyper-V hosts on-premises in a VMM cloud and a secondary VMM server in a 2nd datacenter. Replication between one or more Hyper-V hosts on-premises in a VMM cloud and a secondary VMM server in a 2nd datacenter, using SAN replication. Replication between one or more Hyper-V hosts on-premises in a VMM cloud to a secondary VMM cloud on the same server. In my opinion, it could be interpreted as there is always a business scenario available that can benefit from the features of Azure Site Recovery. 6
6. How Azure Site Recovery Works 6.1 Characteristics Now you have a clear understanding of the different scenarios where Azure Site Recovery can help you in building your disaster recovery plan, it s about time I explain to you how it actually works. While there are a few (minor) differences, based on what source system you are starting from (Hyper- V hosts, Virtual Machine Manager clouds, VMware or physical servers without a hypervisor), overall, the base idea remains the same; allowing for virtual machine system and data information, which gets replicated from the on-premises datacenter to Azure. This replication goes over encrypted https port 443 traffic, so your data is secured in transit. That s already easy to understand. The next part which is also pretty easy, is the configuration itself. As with a lot of Azure components, configuration is mainly wizard-based out of the Azure portal, although PowerShell is also a good alternative for more complex environments or when needing more granular recovery plans. 6.2 Configuration The full ASR configuration can be summarized in 7 steps: Step Step 1 : Create a vault Step 2 : Create a Hyper-V Site Scenario Create an Azure Site Recovery vault. Create a Hyper-V site as a logical container for all the Hyper-V servers that contain virtual machines you want to protect. Step 3 : Prepare Hyper-V Servers Generate a registration key and download the Provider setup file. You run the file on each Hyper-V server in the site and select the key to register the server in the vault. Step 4 : Prepare Resources Step 5 : Create and configure Protection Groups Create an Azure storage account to store replicated virtual machines. Create a protection group and apply protection settings to it. The protection settings will be applied to every virtual machine you add to the group. Step 6 : Enable Protection for VM s Enable protection for virtual machines by adding them to a protection group. Step 7 : Test Deployment Run a test failover for a virtual machine. 7
As is visible in the Azure Recovery Services portal: 7. Outcome Once your on-premises datacenter or host is being protected using the ASR provider and data gets replicated through the Azure Site Recovery, every virtual machine or physical host is configured as a VHD-virtual disk, which gets linked to a virtual machine profile. It is important to note that the virtual machines themselves are not running, as long as the recovery process itself is not initiated. Basically, you are only consuming Azure storage, which makes this solution great from a budget perspective. It s only when the recovery process starts (your VM or VMs boot up), that it will start to cost consumption. When in failover mode (your machines are not available on-premises anymore and running in Azure), you have two possibilities to recover the environment. You either initiate a fail-back scenario from Azure back to the datacenter once the services are restored on that side, or you leave the VM s running in Azure and use Azure as the primary datacenter from now on. 8. Monitoring While Azure Site Recovery provides extensive details on the Azure Recovery jobs being executed in real-time as well as providing historical information, sometimes it is not that easy to consult. This is where I would like to share a few words regarding Live Maps Unity, Savision s business service management solution, which comes with unlimited HTML5 dashboards. First of all, what I see as the main benefit of the solution, is how easy it is to configure and use. Secondly, what makes it a good recommendation is its proven integration with other System Center components like Operations Manager. 8
With minimal effort, executive-level dashboards can be created showing the overall health of all critical business services, including those relying on Azure Site Recovery. Executives have all the information they need at their fingertips, such as overall operational health and SLA compliance without having to open technical tools or query technical teams. 9. Summary With all new features and services coming out in Azure one-by-one at a very high speed, it is sometimes hard to convince customers in taking the steps to adopt Azure. From my own experience, I saw a lot of opportunities and power in the Azure Site Recovery feature. A true disaster recovery solution and a decent cost, fully comparable to your health insurance plan. You pay a certain fee (Azure storage in this case), and only start paying when you rely on the policy (starting VMs as part of the recovery plan in our scenario). Knowing this DR mechanism is supported between two datacenters, where ASR acts as an orchestrator for failover/failback, as well as between on-premises Hyper-V hosts, VMware hosts or physical machines with no hypervisor running, makes it one of the best solutions available in the market today to build your enterprise disaster recovery plan. If you want to know more about Azure Site Recovery, you can consult the Azure documentation here so you can see for yourself what a great solution Azure Site Recovery actually is: More documentation: http://azure.microsoft.com/en-us/documentation/ Register for an Azure trial subscription here: https://azure.microsoft.com/en-us/pricing/free-trial/ 9
About Peter De Tender Peter De Tender is a freelance Microsoft Infrastructure Expert, Microsoft Certified Trainer (MCT), MVP on Windows Expert IT Pro, STEP member and appreciated international speaker at conferences like TechEd US and Europe, NICConf Oslo, TechFuse, Microsoft Ignite Chicago and a lot more. Starting his IT career in 1996 on Windows NT4 and Exchange Server 5.5 platform implementations, Peter has never looked away from Microsoft Server technologies both professionally and in his spare time. Peter is a technical writer for Petri, has authored several books for Packt Publishing, and writer of multiple whitepapers. He is currently writing a book for Apress on Azure Site Recovery, as well as the author of the Azure Site Recovery courseware for Opsgility. Peter's enthusiasm as Microsoft Certified Trainer led him to found the IAMCT Belgian chapter in 2010, as well as taking the position of European Chairman since Jan 2013. His dedication to the IT Pro community and TechEd conferences was translated in founding the TechEdYellowPantsTeam, a small community which goal is assisting IT Pro s in all things Microsoft and certification. For more information or to follow Peter, you can check out his website or follow him on Twitter: @pdtit About Savision Savision is the market leader in business service and cloud management solutions for Microsoft System Center. The company s monitoring and visualizing capabilities bridge the gap between IT and business, by transforming IT data into predictive, actionable and relevant information about the entire cloud and datacenter infrastructure. Savision's intuitive and customizable dashboards provide context for each business service, increasing organizational efficiency, reducing IT operational costs up to 20%, and preventing IT-related problems and business downtime. Savision s solutions scale from small and medium businesses and government bodies to Fortune 500 companies operating in different fields and have been adopted by over 700 organizations worldwide. Savision is headquartered in Amsterdam, The Netherlands, and has offices in Dallas and Ottawa. For more information, visit 10