Hyper V Cluster vs Fail Over Cluster backup?

[a-reaver]

Hi,

I've just finished setting up out S2D setup with a fail over cluster and now in the process of setting up Veeam. Just wanted to find out what would be the best way to go about backing this up.

I see there is a guide about setting up the Fail Over Cluster but it doesnt really explain what benifits it gives me or how i would do restorations. How is it different to Hyper V Cluster backups?

Originally i was going to setup the Hyper V cluster as a backup option but found there isnt really much documentation to it, everything seems to be about backing up the Cluster itself.

All i really want to be able to do is backup the VM's and restore the VM's or Files if there are issues.

I have backup and replication 9.5

[HannesK]

Hello,
the user guide does not have any storage spaces direct specifics because there is nothing specific with S2D. It's just the storage and the software does not care whether it's iSCSI, FibreChannel, local disks or S2D.

As you are new to Veeam, I recommend the quick start guide: https://helpcenter.veeam.com/docs/backu ... ml?ver=100

Best regards,
Hannes

PS: if you are running a new setup, I suggest to start with version 10

[a-reaver]

Hi Thanks for getting back to me.

Im chasing up with our supplier as to how to upgrade our license to v10 as the support contact expired, so stuck wit 9.5 for now.

We currently have it setup and working for the past few years on our 2 Servers running out VM's in Hyper V with no issues. Now with these new servers that have been configured as a failover cluster i have to change to a different style of backing up.

Which is where i am at now. Should i be setting up to backup our server via the Microsoft Hyper V Cluster backup option or do the Active Directory Protection Group for the cluster? I havent been able to find any information about what the difference is between them or which is the right way of doing it.

[HannesK]

Hello,
I do not understand the question. Except for very small environments with only one host, all Hyper-V installations are failover clusters. There is no "protection group" for VMs. This only exists for agents.

Did you follow the quick start guide I mentioned and added the cluster? https://helpcenter.veeam.com/docs/backu ... ml?ver=100

You can do normal VM backup.

Best regards,
Hannes

PS: don't configure off-host proxies. just keep the default. I asked the documentation team to remove off-host proxies from the quick-start guide because they are complicated in real world

[a-reaver]

Hi,

I actually have 2 hosts, running microsoft S2D as a failover cluster. I can setup the Hyper-V Cluster like you suggested, it was how i actually was intending to do it at the beginning but when i started researching online it seems like there was another option which is Backup a Windows Fail Over cluster. https://www.veeam.com/kb2463

the question is which should i using? what is the diference?

[HannesK]

Hi,
yes, every Hyper-V cluster runs as failover cluster. That's normal. No matter which storage technology is used

The KB article you posted is agent based backup of physical servers. But you want VM based backup (as I mentioned earlier)... because if you would backup the cluster, then you would get just a bunch of VHDX files without any proper restore options.

Best regards,
Hannes

[TheOnlyWizard17]

Hi a-reaver,
S2D info is very sparse at Veeam. Even now with v10... Docs are behind as there's only one little mention about it in the chapter about proxies. As S2D leverages LOCAL storage, you can't use a backup proxy to offload the backups to a host/VM/whatever outside of the cluster, meaning that all backups always will be running in On-Host mode, where the "proxy" used is the cluster-node doing the VM backup itself. This means that it is of moderate concern that doing backups frequently or in great numbers (many concurrent tasks) will have an impact on your production-storage or the VMs running on them. So, as far as configuration is concerned you can simply add the cluster and configure a job to either backup the entire cluster, or VMs, as per normal. If you're cluster has RDMA-NICs (as a requirement for S2D anyways) then sadly to say, Veeam v10 has no direct builtin support for this so they won't get used. You can force using at least one of them by specifying the NICs subnet in General options -> network traffic rules->Networks. That way, Veeam will still use TCP (not RDMA) but will do so over one RDMA NIC (That is, provided the recieving side, your VBR backup-server, also has a NIC with that Subnet, preferably also RDMA, so same speed, NIC. If your VBR is a VM on a stand-alone Hyper-V Host with RDMA NICs, as I have, you could use virtual RDMA inside the VBR VM to have it use the same NIC/Subnet)
Because backups are always On-Host mode, and Veeam uses TCP to get the backups across your VBR Repository, expect "heavier than you would expect" type of loads if you use the RDMA NICs, as their speed usually is so much higher, that CPU-load also increases as the traffic doesn't get offloaded to RDMA. If you would want Veeam to use the RDMA NICs AND use SMB/RDMA, there is a "rather complex" workaround for it, but currently it means giving up a few features you normally enjoy, such as auto-New VM inclusion in the Job, but it is possible. I filed a feature request to support how they could make Veeam use RDMA out of the box if it detects a S2D cluster, but we'll have to see if Veeam will approve and will be able to implement it...