2

I need to build a 2 node Hyper-V failover cluster, that's dead simple right? The problem is that the hardware is already bought.

Both servers are identical and approved.

Related spec:

  • 250GB Drive for OS ( Microsoft Hyper-V 2k12 R2)
  • 14TB raid10 drive
  • 64GB RAM and so on.

Plus we have a Synology DS415+ NAS which is great, but my boss wants to use the internal drives as CSV and the NAS will be as a witness disk (with 2TB drives.. sigh)

If is it possible, is there a proper way of doing this? Buying another NAS for clustering the storage is not an option.

At first I was thinking about making both nodes iSCSI Target Server and Initiator simultaneously but I don't think if it is the right thing to do!

Improve this question
7
  • You need to use shared storage for your CSV. Commented Nov 12, 2015 at 14:50
  • So there's no other option than to use the NAS right? than 2x14TB storage is a waste.. Good job by boss for not reading things up :) and thank you for your reply! Commented Nov 12, 2015 at 17:10
  • 1
    My apologies, I didn't get very specific in my comment. I didn't say that you couldn't use the local storage, I said that you need to use shared storage. You can use the local storage as shared storage if you use something like a Virtual SAN that can create shared storage from the local storage on each host. Note that this adds another layer of complexity to your design, but it is possible. Take a look at this as an example - starwindsoftware.com/starwind-virtual-san Commented Nov 12, 2015 at 17:47
  • 1
    Thanks for the more in-depth answer :) I was looking at Starwind's product in the afternoon. We will probably give it a go and find out if it suits our needs or not :) Big thanks to you anyway! :) Commented Nov 12, 2015 at 18:38
  • Glad to help... Commented Nov 12, 2015 at 18:46

1 Answer 1

Reset to default
0

You can form a cluster using the NAS. But if you want to make the VMs highly available, which is what most people want when they create a Hyper-V cluster, you'll have to store the VMs themselves on the NAS, ignoring the local RAID controller and all the storage connected to it.

CSV is a technology for taking disks that are visible from all nodes of the cluster and making the file system on those disks symmetrically available from every cluster member. The local RAID controllers aren't symmetrically available from every node of your cluster. They're only available to the node that they sit within.

I don't know much about your NAS. If it supports SMB3 or iSCSI, it can be the storage for the cluster. The problem is that, unless the NAS itself is itself highly available (multiple power supplies, multiple NICs, multiple compute nodes, etc.) then it's a single point of failure. And if you're willing to tolerate a single machine failure, you probably don't want a Hyper-V cluster at all. Just put the VMs on the individual hosts. They'll fail if the machine that they're running on fails. But if you use a non-redundant NAS, your VMs will fail if either the host fails or the NAS fails.

Improve this answer
1
  • We don' want to have any point of failure. I have already set up a NAS cluster, but the problem with that one is that the NAS can fail too, and it has a limited ( 4 bay) storage capacity. It supports iSCSI and has a built in NAS-CLUSTER if we buy an other NAS. Right now i'm testing Starwind's Virtual SAN but it is really pricey for an SMB like we are, but it works too! And the real problem is that the hardware is already bought and I can't change it, so if we use NAS or SAN then the 2x14TB storage will be wasted for nothing. Commented Nov 16, 2015 at 12:33

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.