Hurricane Matthew – Start Those Planned Failovers

A hurricane is about to blast it’s way up the east coast of the USA, making landfall in south Florida probably early on Friday morning, and working it’s way up to Norfolk, VA, by Sunday morning. We know how much damage these hurricanes can do, especially if tides rise and seawater starts mixing with electric, servers, and storage – we’re talking not just business down, but business offline, and maybe even business dead. I’m sorry, but even a stretch cluster to a nearby location is subject to the same mess.

This is when a true DR solution is required. “But I cannot afford a DR solution”, you say. You can’t afford to not have one, but I do know what you could have deployed (it’s too late now, by the way, if you are in the target zone for Hurricane Matthew). Azure Site Recovery (ASR) is an OPEX-based way to get a DR site in the cloud. The cost is a monthly drip feed instead of the CAPEX big bang that a traditional DR site is:

  • $25 per replicated machine per month, in Azure South Central US.
  • Replicated disk storage starts at $0.05 per GB in the same Azure region.

The solution works with:

  • Hyper-V
  • vSphere
  • Physical servers

And it’s really simple to use and reliable; thousands (if not more) of businesses are deploying and testing ASR failovers on a regular basis. This out-of-“the box” shared platform is tested constantly, which makes it way more reliable than some home-baked solution.

You get full orchestration – so if I saw the forecast today, I could start my business continuity plan, start the failover and hit the road. My machines would start a planned failover (ordered and no data loss) to Azure and would be waiting for me when I get to my rendezvous point. Note that my orchestration can also kick off PowerShell scripts (Azure Automation) to do some fancy things, such as redirecting internet traffic that I had routed using Azure Traffic Manager.

If you have ASR and are in one of the areas that will be affected, then do a test failover, do any required remediation’s, and then start that failover. Hopefully, your business is not damaged and you can do a failback afterwards (if you want to). If you don’t have a DR solution, I hope you survive, and have the sense to look at ASR soon afterwards – it is hurricane season!

Technorati Tags: ,,

Future Decoded: My Session Is “Azure Site Recovery – Be A Super Hero!”

I’m going to be talking about Azure’s DR-as-a-Service or DR-site-in-the-cloud solution, Azure Site Recovery (ASR) at Future Decoded, a fantastic IT event by Microsoft UK beside London City Airport, on November 1/2.

“Remember; when disaster strikes, the time to prepare has passed” , Stephen Cyros.

We all think that disasters never happen near us; bushfires, earthquakes and flying cows are things that happen elsewhere. But the truth is very different, disasters strike every day without making headlines, sometimes wiping out a company or just that one critical server, and the cruel thing about disasters is that they tend to strike those that are unprepared; it’s those times that the business needs a hero. Unfortunately, a hero needs to be prepared, and during a disaster is not the time to prepare. IT Pros know that we need to have DR solutions, but often they’ve proven to be too costly or too difficult to implement. Times have changed; cloud computing has democratized and simplified DR. ASR’s low cost OPEX model makes replication of physical, vSphere, or Hyper-V servers to Azure more … more so now, thanks to recent price cuts. Large and small enterprises benefit from ASR’s orchestration which makes failover easy and reliable – you can order failover of machines and build in scripted extensions, and test your orchestrated failover without impacting production systems.

TW_FD-Register-banner_1024x512px (002)

Future Decoded will have lots of great content from a variety of speakers with different backgrounds, and come along to my session to learn how you can be the super hero, and get your business back operational when everyone else is panicking.

New Features in Azure – 29 April 2016

Microsoft sent out an announcement about feature changes/additions in Azure last night. Some news there that was much quicker than I expected …

Azure Container Service Generally Available

Azure Container Service helps customers manage container-based applications in production, at scale. Azure Container Service is fully integrated with the Azure portal, Azure Resource Manager, and our compute, storage, and networking resources. This supports Docker images, using familiar tools and either open-source DCOS or Docker Swarm as the underlying orchestration technology. The only cost for Azure Container Service is what you pay to use the underlying resources.

New Azure Storage Cool Tier Generally Available

Azure storage was cheap already, but it just got cheaper. Now we have a lower tier for blob storage, that can be used for services like backup, disaster recovery, or data archival. As a result, the storage pricing page has been updated to reflect the new options. Here is the pricing for block blobs (backup) in North Europe:

image

LRS, what we typically see being used for on-prem backup, costs (in North Europe):

  • €0.0084 or $0.01 per GB in cool storage
  • €0.0202 or $0.024 per GB in hot storage

Seriously, that is cheap. Microsoft has detailed the transaction pricing too (this really only affects huge deployments):

image

As you can infer from the above, cool storage really is for infrequently accessed data. Hot storage is where frequently accessed data should reside.

The SLA on cold storage is lower – at 99% which is still pretty damned good, especially for the price. You can bump that up with the RA-GRS redundancy option, where Microsoft provides a higher read SLA of  99.9% for the Cool access tier.

image

Note that I don’t yet see a way to use cool storage with Azure Backup, in either the old or the new portals. But a number of 3rd party backup tools can use it. Note that you have to create a new storage account type called BlobStorage in ARM to gain access to hot/cold, and you can convert from cold to hot, and vice versa (the above transaction costs will be charged for conversions starting on June 1st).

Azure Site Recovery Portal General Availability

The DR solution, ASR, now has support in the new Azure Portal. This adds support for ARM (CSP). Improvements include:

  • Azure Resource Manager support for all scenarios
  • First-class support for Cloud Solution Provider (CSP) subscriptions
  • Streamlined Getting Started experience for all Site Recovery scenarios
  • New Policy construct for flexible association and management
  • Functionality of Backup and Azure Site Recovery in a single vault construct

Enhanced VMware to Azure:

  • New Exclude Disk functionality when replicating VMware VMs to Azure
  • Support for Premium Storage for high-churn workloads

I still don’t see ASR as being ready in ARM. Yes, I can replicate and failover VMs, but I see Azure AD and RemoteApp as essential pieces to the solution. What good are machines in the cloud if I cannot access them? Yes, I can use point-site VPN (don’t assume you’ll have site-site VPN option from your alternative office, e.g. a hotel meeting room) and the “fun” that will cause, but how will my legacy applications perform … after I’ve spent hours installing them on new laptops I just bought? Unfortunately, Service Manager (Azure V1) continues to be my recommendation for DR solutions in Azure, and the rumblings tell me that it’ll stay that way for another 6 or so months 🙁 It’s a pity because the new ASR UI is pretty nice.

Technorati Tags: ,,

Azure Global Bootcamp Dublin – When Disaster Strikes

I spent Saturday afternoon in the offices of Microsoft Ireland at the very successful Azure Global Bootcamp event in Dublin. Other speakers covered a variety of topics for the 160 (approx) attendees and I wrapped up the day with a session on using Azure Site Recovery as a virtual DR site in the cloud for Hyper-V, VMware, and physical servers.

I was pretty exhausted going into the session, but it was good fun for me to do it. The crowd was engaged, and they even laughed at one or two of my attempts at humour. There was loads of engagement afterwards which was as much fun, even if maybe 95% of the audience were developers Winking smile

You can find my PowerPoint deck on SlideShare:

Here are a few photos that some folks took:

Media preview

Starting off [Image credit: Niall Moran, Microsoft]

One of the two rooms used on the day [Image credit: Ryan Mesches, Microsoft]

image

I stood between the audience and food – so I had some fun [Image credit: Rob Allen, Unity]

Media preview

Vikas Sahni (organiser & speaker), Bob Duffy (SQL MVP and speaker), and me.

About 95% of the audience identified themselves as developers to one of the previous speakers. Around 40% of the room claimed to already have DR services in place. So I’m curious why so many stuck around for an IT pro topic on DR. Maybe they wanted a cheaper, cloud-based alternative?

Global Azure BootCamp 2016 – Dublin

Microsoft and “the community” are partnering once again to run The Azure Global BootCamp. ICYMI, the boot camp is a one-day event in locations around the world, where Azure veterans share their knowledge with attendees at this free event.

This event is running in Dublin at 09:30 on Saturday April 16th at Microsoft Atrium Building B, at Carmanhall Road, in Sandyford Industrial Estate, Dublin 18.

The agenda is:

  • What’s new in Azure – Niall Moran (Microsoft)
  • Building and Deploying Azure App Services – Aidan Casey (MVP)
  • Migrating SQL to Azure, an Architectural Perspective – Bob Duffy (MVP)
  • Building Real World applications – Vikas Sahni
  • When disaster strikes – Aidan Finn (MVP)

My session will be focusing on the hybrid cloud solution where Azure acts as a DR site for your on-premises servers (physical, VMware, or Hyper-V).

The event page, with agenda and registration can be found here.

How To Force Azure Replication To Stop From Orphaned Hyper-V VM

There is a scenario when you are using Azure Site Recovery and a VM somehow becomes orphanened, no longer controlled by ASR, but you cannot remove replication from the VM on the host. I had that situation this morning with a WS2012 R2 Hyper-V VM (no VMM present).

The situation leaves you in a position where you cannot disable replication on the VM using either the UI or PowerShell, because the host continues to believe that replication is managed by Azure, even if you remove the provider (agent) from the host or remove the host from ASR. In PowerShell, you get the error:

Operation not allowed because the virtual machine ‘<name>’ is replicating to a provider other than Hyper-V”

Failed ASR Removal

Microsoft has guidance on how to clear this problem up for Hyper-V to Azure and VMM to Azure replication, which I found by accident after a difficult 30 minutes! The key for me to the solution was to run a small 4 line script that removes replication using WMI, found under the heading “Clean up protection settings manually (between Hyper-V sites and Azure)”. I copied that script into ISE (running with elevated admin rights) and replication was disabled for the VM.

The Genuine Need for Disaster Recovery In Ireland/EU

How many times have you watched or read the news, saw some story about an earthquake, hurricane, typhoon, or some other disaster and think “that will never happen here”? Stop kidding yourself; disasters can happen almost everywhere.

I’ve always considered Ireland to be relatively safe. We don’t have (anything you’d notice) earthquakes, typhoons, or tornadoes; our cattle and sheep don’t need flying licenses. Our weather is dominated by the gulf stream, keep Ireland temperate. It doesn’t get hot here (we are quite northerly) and our winters consist of cloud, rain, and normally about half a day of snow. We get the tail end of some of those hurricanes that hit the east coast US, but there’s not much left by the time they reach us – some trees get knocked over, some tiles knocked on our roofs, but it’s not too bad. Even when we look at our neighbours in England, we see how their more extreme climate causes them disasters that we don’t get. Natural disasters just don’t happen here. Or do they?

The last month or so has revealed that to be a lie. Ireland has been battered by 6 storms in the past month. The latest, Storm Frank, was preceded with warnings that the country was saturated. That means that the ground has absorbed all of the water that it can; any further rainfall will not be absorbed, and it will pool, flow, and flood.

This morning, I woke to these scenes:

image

Enniscorthy, Co. Wexford [Image source: Paddy Banville]

Embedded image permalink

Graignamanagh, Co. Kilkenny [Image source: Graignamanagh G.A.A]

image

Middleton, Co. Cork [Image source: Fiona Donnelly]

Frank isn’t finished. It’s still blowing outside my office and more rain is sure to fall. There are stories of communities being evacuated to hotels, and the above photos are just the easy ones for the media to access.

This isn’t just a case of cows trapped in fields, stick a sandbag on it and you’re sorted, or somewhere far away. This is local. And Ireland is a relatively safe place – we’re not Oklahoma, a place that some deity has decided should be subject to cat 5 tornadoes every time you’re not looking. Dorothy, the point is, that disasters happen everywhere, including in the EU where we think it safe.

Let’s bring this back to business. Businesses have been put out of action by these floods. Odds are any computers or servers were either on the ground floor or in the basement. Those machines are dead. That means those businesses are dead. They might be lucky enough to have tapes (let’s leave that for another time) stored offsite but how reliable are they and will bare-metal restore work, or will it take forever? How much money will those businesses lose, or more critically, will those businesses survive loss of customers?

This is exactly why these businesses need a disaster recovery (DR) solution. There are several reasons why they don’t have one now:

  • Fires and other unnatural disasters happen everywhere
  • They couldn’t afford one
  • The business owners didn’t think there was a need for one
  • Some resellers didn’t think there was demand for one so they never brought it up with their customers

The need is there, as we can clearly see above. And thanks to Microsoft Azure, DR has never been so affordable. FYI, it comes in at a price that is a small fraction of the cost of solutions from the likes of Irish companies such as KeepITSafe – I’ve done the competitive pricing – and it opens that customer up to more technical opportunities with hybrid cloud solutions.

Microsoft Azure Site Recovery Services (ASR) is a disaster recovery-as-a-service (DRaaS) or cloud DR site offering from Microsoft. The beauty of it is that it’s there for everyone from the small business to the large enterprise. It works with Hyper-V, vSphere or physical machines, and it works with Windows or Linux as long as the OS is supported by Azure (W2008 R2 or later on the Windows side).

Note: There is a cost overhead for vSphere or physical machines to allow for on-premises conversion and forward and in-cloud management and storage, so you need a certain scale to absorb that cost. This is why I describe ASR as being perfect for SMEs with Hyper-V and mid-large companies with Hyper-V, vSphere or physical machines.

If I had ASR in place, and I has a business on the quayside in Cork, near the Slaney in Enniscorthy, or anywhere else where the rivers were close to bursting the banks then I would perform a planned failover, requiring about 2 minutes of my time to started a pre-engineered and tested one-click failover. My machines would shut down in the desired order, flush the last bit of replication to Azure, and start up the VMs in the desired order in Azure, and my machines and data would be safe. I can failback to new equipment or stay in Azure if the disaster wipes out my servers. And if that disaster doesn’t happen, I can easily failback to new equipment, or choose to stay in Azure and not worry about local floods again.

Technorati Tags: ,,,,,

Microsoft News – 30 September 2015

Microsoft announced a lot of stuff at AzureCon last night so there’s lots of “launch” posts to describe the features. I also found a glut of 2012 R2 Hyper-V related KB articles & hotfixes from the last month or so.

Hyper-V

Windows Server

Azure

Office 365

EMS

Microsoft News – 28 September 2015

Wow, the year is flying by fast. There’s a bunch of stuff to read here. Microsoft has stepped up the amount of information being released on WS2016 Hyper-V (and related) features. EMS is growing in terms of features and functionality. And Azure IaaS continues to release lots of new features.

Hyper-V

Windows Client

Azure

System Center

Office 365

EMS

Security

Miscellaneous

Event: Taking The “Disaster” Out of “Disaster Recovery”

I’m going to be presenting another webcast for Petri.com, sponsored by Infrascale. In this event we’ll be talking about disaster recovery, how you can do it yourself, and how you can leverage cloud services, i.e. Disaster Recovery-as-a-Service (DRaaS).

The live webcast runs for an hour, starting at 13:00 EDT (18:00 UK/Ireland time).

image

Technorati Tags: ,,,