whatif Hybrid Storage Spaces Direct #S2D and Storage Replication #SR #Azure Windows Server 2016 #windowsserver2016

As you know in Windows Server you can use the local storage in you cluster and you can also replicate storage between two Servers.

But what if we combine these two options ? seems logical and with some advanced config it could work. ok but what if I use Azure for this and better I use a Hybrid config using a Cluster that has also a leg in a on premise  will this work ? I don’t think it is supported but cluster validation passed on this so it must be supported. with a *

Ok what Do I need for this :

Azure Subscription – Check

Azure Site – to – site VPN  – Check

on-premise Cluster – Check  6 node Cluster

Azure Cluster nodes – Check 2 nodes running in Azure.

Fast Internet line – check

I’m not showing you all the details else it would be a very long blog post and I have already posted on how you can build your Replica and How to use the Storage Spaces Direct Combine them with those Two options.



My Setup is 4 Cluster nodes On Hyper-v on premise and Two node in Azure all running All Running Windows Server 2016.

and basically what I did is Building a cluster with Storage Spaces Direct based on 3 local disk and on top of this I created a 2 Disks that I used for replication.


as you can see I have 26 disk in Node 1 Different Sizes and Shared and non shared disks these are running on my Hyper-v 2012R2 Server


and with the Storage Spaces Direct option my Cluster Would look like this.  Hybrid Cluster with all the Best options in Windows Server 2016


My Storage Pools One is running in Azure and On is on premis As for the Replication all replication disk needs to be at the same size.


Sizing difficult in azure So I created first the Azure disk to see what size they are and after that I created the On-premise disks


This is really nice all native Windows Server 2016. the only thing you need is a fast internet line. and Currently the limit is the access to Azure ( if you don’t have Expressroute. )


Seting up the Replication is easy with powershell

New-SRPartnership -SourceComputerName win2015-1 -SourceRGName Azure_group01 -SourceVolumeName u: -SourceLogVolumeName v: -DestinationComputerName win2015-6 -DestinationRGName Azure_group02 -DestinationVolumeName p: -DestinationLogVolumeName Q: -LogSizeInBytes 1gb


My Replication with Storage Spaces Direct in a hybrid Configuration. as you can see with this more configurations are possible and if you make sure the line latency is below 50ms things could work just fine.


But building this is fun but imaging you need to trouble shoot this where are my disk and what is failing things are getting complex and even with the hybrid model extending your Datacenter to Azure is a bit closer.


With todays Fast internet it is easy to build this and hybrid solutions are easy to build especially with windows server 2016 build in Replica and local Storage for clustering. extending your lab or production to Azure and you can demonstrate the high available solutions today. but keep in mind building an troubleshooting can be a pain as environments are getting complex.


Checkout the MVP V-Conf Session

Deploying Highly Available SQL Server in Microsoft Azure IaaS



Download Windows Server Technical Preview evaluations:

Happy clustering

Robert Smit



Author: Robert Smit [MVP]

Robert Smit is Senior Technical Evangelist and is a current Microsoft MVP in Clustering as of 2009. Robert has over 20 years experience in IT with experience in the educational, health-care and finance industries. Robert’s past IT experience in the trenches of IT gives him the knowledge and insight that allows him to communicate effectively with IT professionals who are trying to address real concerns around business continuity, disaster recovery and regulatory compliance issues. Robert holds the following certifications: MCT - Microsoft Certified Trainer, MCTS - Windows Server Virtualization, MCSE, MCSA and MCPS. He is an active participant in the Microsoft newsgroup community and is currently focused on Hyper-V, Failover Clustering, SQL Server, Azure and all things related to Cloud Computing and Infrastructure Optimalization. Follow Robert on Twitter @ClusterMVP Or follow his blog https://robertsmit.wordpress.com Linkedin Profile Http://nl.linkedin.com/in/robertsmit Robert is also capable of transferring his knowledge to others which is a rare feature in the field of IT. He makes a point of not only solving issues but also of giving on the job training of his colleagues. A customer says " Robert has been a big influence on our technical staff and I have to come to know him as a brilliant specialist concerning Microsoft Products. He was Capable with his in-depth knowledge of Microsoft products to troubleshoot problems and develop our infrastructure to a higher level. I would certainly hire him again in the future. " Details of the Recommendation: "I have been coordinating with Robert implementing a very complex system. Although he was primarily a Microsoft infrastructure specialist; he was able to understand and debug .Net based complext Windows applications and websites. His input to improve performance of applications proved very helpful for the success of our project

9 thoughts on “whatif Hybrid Storage Spaces Direct #S2D and Storage Replication #SR #Azure Windows Server 2016 #windowsserver2016”

  1. Because SR is configured with drive letters – does Storage Replica still work, when the owner node of S2D goes down?

  2. That is a great question. S2D down is also a wide statement when is this down ? when all disk are gone and the storage pool is no longer available then all the resources on this pool is also unavailable. and yes also SR you could use SR on the DR side 😉
    Therefor you have more uptime. But is this blog case it is all real basic setup but sure in real life you could create more uptime of use this for DR.

    In your question : if the storage pool is offline then there is no Storage Replication As I use the Pool disk for the Replication.

    Thanks for visiting my blog 😉

  3. Hey Robert,

    Thanks for answering. Have you been able to setup SR on a S2D-CSV with Failoverclustermanager?
    I couldnt get it to work with Failoverclustermanager.

    It would be nice, if the SR-Setup could be somehow clustered, so that SR does still work, even when the node (just one node) on which it is configured fails.

  4. You can’t create two pools in one cluster as the disk are claimed by the node as soon as you enabled this. But you can create multiple disk on this pool. and be flexible on how to use this pool. at the end it all sits on the same disks but you can.

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: