Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Question
Tuesday, December 19, 2017 9:11 PM
In preparation for my MCSA 70-412 exam, I have purchased a used 2U server, and am trying to become familiar with failover clustering.
I am using Hyper-V as my hypervisor, and am working with two different VMs running Windows 2012R2, with the File Server role and features installed. These machines are named MCD-FS1, and MCD-FS2. I then installed the Failover Clustering feature for both of these VMs.
But, creating a failover cluster with the File Server role requires cluster storage. I do not have an iSCSI SAN available in my home lab, so I have attempted to create several disks for use in the cluster storage by creating VHDs with my hypervisor. I then added the VHDs to MCD-FS1 in the VM's settings in Hyper-V.
I connected to MCD-FS1, and I used Disk Management to bring the disks online. I did not create a volume on them. The disks were then selectable to add to cluster storage. But, none of the disks are able to come online.
Solutions on similar TechNet forum posts and google searches do not seem to apply, as these are referring to antivirus software or iSCSI hardware vendors.
I attempted to create basic volumes on the disks through Disk Management and assign them drive letters, but that did not change the outcome.
All replies (6)
Wednesday, December 20, 2017 6:24 AM ✅Answered
Hi josephfmcd,
> I do not have an iSCSI SAN available in my home lab, so I have attempted to create several disks for use in the cluster storage by creating VHDs with my hypervisor.
Are the disks local disk that only available to one node? If yes, these disks are unable to be used by cluster.
For 2012R2 cluster, we need to use shared disks which are available for all nodes to access.
In a test lab, we may need an additional storage server and install ISCSI target server:
In this server, new virtual disks for the cluster nodes to connect:
On the cluster nodes, connect to the ISCSI virtual disk:
Then, in the cluster node's disk management, online and initialize the disk and assign volume to the disk. Online the disk in all nodes while just need to initialize on one node. Then we may add the disk into the cluster for cluster use:
On server 2016 cluster, we may use local disks for cluster use, which is S2D technology.
Best Regards,
Anne
Please remember to mark the replies as answers if they help.
If you have feedback for TechNet Subscriber Support, contact [email protected].
Wednesday, December 20, 2017 6:41 AM
Hi,
Have you selected the "share this VHD" option in the hard disk advanced features?
After that, add the same disks (VHDs) to the secondary node with the option checked.
Please post the results here.
Ideally this is not recommended but since it s a lab you can go ahead with it.
Also you can try the option that Anne has mentioned
Regards,
Bala
Wednesday, December 20, 2017 1:14 PM
The simplest way to create VM clusters is to use the shared VHDx feature. This was a feature introduced in Windows server 2012 R2 that allows the shared storage requirement of a failover cluster to be satisfied by using simple VHDX disks. No need to set up a separate iSCSI environment. No need to set up a shared file system.
By default, this capability is supported in any physically clustered environment. It is not supported in a single stand-alone physical host, but it is very easy to implement and it works fine for a lab environment like you are describing.
The ability to share VHDX files among VM node members is dependent upon the ability of the host environment to shared the VHDX files. This requires a special filter driver that is installed as part of creating a cluster on the physical hosts. As you have just a single host, you need to install the failover clustering feature on the physical host, but you do not need to create a cluster on the physical host. You just need access to the special filter driver. So install the failover cluster feature and then issue the following command:
fltmc.exe attach svhdxflt f:
This will attach the physical drive f: to the filter driver. Store the VHDX files to be used as shared storage on drive f:. You will need to issue this command each time you start the physical host because it is not loaded automatically on systems that are not members of a failover cluster.
Once you have attached the drive to the filter driver, you then have the option in the Hyper-V Manager to define a VHDX to be used as shared storage.
Again, not supported, but for a lab environment this works quickly and easily. Much easier than trying to set up a separate iSCSI or SMB environment to use for shared storage. Of course, if you want to gain additional experience in failover clustering options, definitely try building the other environments. I would recommend first learning about SMB storage for clusters and then moving to iSCSI. Though iSCSI has been around for a while, it is not the direction in which Microsoft is moving for shared storage.
tim
Thursday, December 21, 2017 1:43 AM
Hi josephfmcd,
These are several methods provided above to prepare disks for cluster use, if these replies are helpful, you may mark them as answer, if you have any other concerns, welcome to feedback.
Best Regards,
Anne
Please remember to mark the replies as answers if they help.
If you have feedback for TechNet Subscriber Support, contact [email protected].
Thursday, December 21, 2017 6:27 AM
The options mentioned above are completely great and you can use the one that seems quickest to you but since the shared disk/VHDX or iSCSI LUN are located on a single host there is no way for you to test how quick migration works in case of a host failure. In order to test this feature (which is most probably required for an exam) you can use free StarWind https://www.starwindsoftware.com/starwind-virtual-san-free that works exactly the way Anne already described but creates a mirrored shared storage on both hosts participating in the cluster (and remains online even if one of the hosts fail) that allows you to play with Microsoft Failover Cluster high-availability and fault-tolerance features.
Hope this helps.
Best regards,
Taras Shved
StarWind Software
Blog: Twitter:
LinkedIN:
Note: Posts are provided “AS IS” without warranty of any kind, either expressed or implied, including but not limited to the implied warranties of merchantability and/or fitness for a particular purpose.
Thursday, December 21, 2017 1:21 PM
"creates a mirrored shared storage on both hosts participating in the cluster"
Except the original poster stated he has only a single physical server, not two. But you are correct, otherwise, that he will be unable to try out certain features due to his lab configuration.
tim