Backing up a Windows Failover Cluster with Shared Storage?

[simonpooley]

Hi guys,

Sorry for bumping an old thread but this is exactly what I'd need to achieve. I'm hoping perhaps this ability has been added in a newer version of Veeam?

Thanks

[ChrisGundry]

Hi Simon,

I asked the same thing a few weeks ago but the gist is that Veeam are unable to do anything because VMware cannot snapshot the VM/disks. I asked if there was any suggestion about a way forward, perhaps using Veeam Endpoint within the guest, or some other guest based component but the answer was a flat no. My topic was also merged with another pre-existing topic with no notice or explanation. So unfortunately I expect you will get the same response.

Chris

[Vitaliy S.]

Hi Simon, please review this existing topic for more information, but Chris is correct that in order to backup these VMs you need to take a snapshot, which is not possible to do in current version of VMware vSphere.

[ChrisGundry]

Vitaliy: You said this previously:
To the best of my knowledge, some people are disabling bus sharing during backup job or reconfigure their infrastructures to be supported for VM snapshots.

Can you elaborate please?

[Vitaliy S.]

Yes, sure. If they want to make a backup, they re-configure the VM to allow snapshots, not sure how automatic this process is though.

[simonpooley]

Hi Chris, Vitaliy,

Thanks for responding to me succinctly.

The reason I ask is that we've just created a series of VMs for an application called BizTalk. It works best with a Windows Cluster (2 nodes) because this gives us high availability on some windows services. We also have shared disks on these two clustered application servers.

Because we've moved from old physical servers to VMs in HyperV 2012 (sorry I now realise this is a VMware board but I suspect the issue is the same), I was naively under the impression taking a backup of the machine(s) would be easy! Imagine my surprise when my infrastructure guys are telling me "we can't backup clustered machines at all".

So basically I was looking for a mechanism where we can take periodic snapshots of a two machine clustered VM which uses a SAN for storage. If anyone has any suggestions on how we could go about that I'd be really grateful.

Many thanks

[Vitaliy S.]

Simon, are you using shared VHDX for this VM configuration?

[simonpooley]

Is a shared VHDX the same as a Cluster Shared Volume that is mounted in the C: drive, for instance:

C:\ClusterStorage\DiskName1
C:\ClusterStorage\DiskName2

Our clustered application servers are *not* set up like this (however the db servers are).

I believe our application servers are using shared storage via an iSCSI interface. These drives appear as standard drive letters on the active node, i.e.
Node1 (Active) - C:, D:, E:
Node2 (Inactive) - C:

[dellock6]

Wait, so in this case you can backup (at least in crash consistent mode) the shared VHDX, the limit is about VMware and because it was in this subforum we were a little bit misleaded....

[simonpooley]

Oh ok, so in HyperV a periodic backup of the shared VHDX is possible? That would be great. I'll see if we can do some lab testing.

Thanks

[Vitaliy S.]

Yes, that's possible. Please continue posting in this thread (Hyper-V subforum) if you have any other questions.

[ChrisGundry]

@Vitaliy.
But surely when they 're-configure the machine' (you have not really elaborated on what you mean by that), the machine cannot be a cluster server anymore because it must have bus sharing turned off to perform snapshots. So this method isn't really a possibility is it?

[foggy]

Vitaliy meant any re-configuration that would allow taking a snapshot of the VM.

[frederic.guyon]

I've read all the Forum about difficuties to saved a Windows Failover cluster on a 2 nodes active passive or more nodes active-active-passive !

And I saw that all have the same problem as me with that configuration mainly due to shared storage and special ISCSI controler define in VM. I have no solution but after my different tests I just whant to share my proposition the the Veeam B&R developer team.

Could the new release can process Windows cluster as a type of backup using Windows api like Windows update do the process to update cluster (Cluster AWARE Updating). Not necessary to transfert all roles on a node during the backup but probably a mixed solution with END POINT Backup and VEEAM B&R, but the END POINT Client should be remote controled by VEEAM Proxy.

VEEAM B&R is the probaly best solution actualy on the Market to Backup VM but surely the best one on restore VM instantly. I think the if VEEAM what to continue to be the BEST one it should integrate Cluster backup features, Those kind of Config will be present more and more and Veeam should do something on that.

Thank you for your feed back and sorry for my English, it's not my natural language!

[k1ngb0ra]

Hello community,

I would like to implement a sql-cluster deployment with cluster shared volumes which requires shared disk.
Is there any info/guideline how to implement a veeam-backup in this scenario because there is still the requirement for physical mode raw device mapping?
But physical mode does not support snapshots.
And virtual mode is only supported for cluster in a box (CIB).

any feedback would be greatly appreciated.
Tanks!

[Vitaliy S.]

Hi Boris,

The best practice would be not to use bus sharing, as snapshots are not supported for these kind of disks. Usually people use SQL Server AlwaysOn instead of clustering it via Failover Cluster feature. Please review this topic for more info.

Thanks!

[lando_uk]

I was testing Server 2016 Storage Replica to possibly replaced some old DFSR, but unfortunately that also has to use bus sharing so my plans have been scuppered

[k1ngb0ra]

Hi Vitaliy,

thanks for your reply, but unfortunately AlwaysOn is not the same as a "real" FCI, in my opinion AlwaysOn is like a narrow gauge cluster without the need for shared storage.

An FCI covers everything about the SQL instance - databases, logins, certificates, policies, agent jobs, and more. Whether you have two databases or hundred databases, they will all go together. An AvailabilityGroup will only fail over the data in the databases that are in the group. You must separately manage everything else. The more replicas you add, the more complex management becomes.

maybe this is one reason I'm considering to move away from VMware to Hyper-V because there are snapshot capable shared VHDX available

[Vitaliy S.]

Ok, thanks for clarification, makes sense. If you're considering VHDX, then please take a look at this topic > Backing up a Windows Failover Cluster with Shared vhdx?

[jguillot]

Hello,

I would like to use Microsoft Failover Cluster with 2 Windows Server 2016 servers to make high availability file servers with shared volume for data.

Here is my architecture:

Main site:

• 4 ESXi in connection (iSCSI) with HPE StoreVirtual VSA for storage (SDS).

Emergency site (DR):

• 2 ESXi in connection (iSCSI) with a big NAS.

Backup infrastructure:

• Veeam Backup & Replication

• HPE StoreOnce VSA

• Main backup on a DAS Bay

• Backup copy on a remote NAS

My goal :

• The 2 file servers would be on ESXi different from the main site.

• Shared storage of file servers would be on a volume of the HPE StoreVirtual VSA.

• Backup the VMs on the DAS array (through Veeam B&R and StoreOnce VSA) using StoreVirtual VSA Storage Snapshots.

• And also to continually replicate all my VMs to my DR site.

1) Is it possible?
2) Shared storage of file servers can be a VMDK? At first I thought I would use vRDM but apparently Veeam does not support Storage Snapshot on them. No worries with vMotion?
3) How is the backup of the 2 file servers with a shared VMDK? the VMDK is saved 2 times?
4) How is replication done with a shared VMDK?

thank you in advance

Ps : Sorry for my poor english

[foggy]

Veeam B&R itself doesn't support VM's with shared storage, so I'd look at Veeam B&R 9.5 Update 3 that adds Windows Failover Cluster support for Veeam Agent for Windows.

[simon.kobben]

Veeam B&R itself doesn't support VM's with shared storage, so I'd look at Veeam B&R 9.5 Update 3 that adds Windows Failover Cluster support for Veeam Agent for Windows.

I read the release notes of Update 3. Does this mean, that all the limitations from KB2304 https://www.veeam.com/kb2304 are gone?
Or is there a detailed documentation about what is supported?

[Vitaliy S.]

Hi Simon,

Not all limitations are gone. Please review this and this topic in the Agent forums. Thanks!

[albertwt]

Is there any update for backing up the SQL Server cluster Always on MSCS (VMware ESXi) ?
because at the moment to avoid Database failover, I'm installing Veeam Agent backup and treat the VM as a physical server.

[Vitaliy S.]

No changes in that regard.