Backup Server with NVDIMM

[GianlucaCroci]

Hello,
we are using B&R 9.5 Update 3a and we backup 4 physical machines with "Windows Server 2016". We've Clustered these machines, and we've installed SQL 2016.
We backup them as a "Failover Cluster", we use as "Name Computer" the name of the Cluster, and the "Entire computer" as "Backup Mode" (so we don't have to give all the disk that are used).
Now we want to use NVDIMM Memories (formated as DAX), because we want to increase the speed of the logs.
Added and configured this Memories and when we run the backup job, it fail (message below).
We put it offline, and the backup was good.
We've changed the access rights to this Disk, excluding the user who performs the backup, thinking that not being able to access it, this was not taken into consideration by the job, but instead the job is still in error and we get the error message next (same as before).

"Creating VSS snapshot Error: Failed to create snapshot: Backup job failed. Discovery phase failed.
Cannot add volumes to the snapshot set. Cannot add a volume to the snapshot set.
Volume name: [\\?\Volume{d0b0be00-6a57-4c9c-b6b7-2bc73b011269}\]. Cannot check whether volume supported by the system software provider.
VSS error: VSS_E_UNEXPECTED_PROVIDER_ERROR. Code:0x8004230f
Error: Failed to create snapshot: Backup job failed. Discovery phase failed. Cannot add volumes to the snapshot set.
Cannot add a volume to the snapshot set. Volume name: [\\?\Volume{d0b0be00-6a57-4c9c-b6b7-2bc73b011269}\].
Cannot check whether volume supported by the system software provider. VSS error: VSS_E_UNEXPECTED_PROVIDER_ERROR. Code:0x8004230f"

What's wrong? Any idea how we can proceed without having to completely redo the job?

I repeat, this is a SQL Cluster with nodes that possibly move between one Host and another, and each with multiple associated disks.

Thank's for the help.
Kind regards

[HannesK]

Hello,
looks like that VSS does not work on the DAX volumes. Did you try to manually create a VSS snapshot on that volume?

Best regards,
Hannes

[GianlucaCroci]

Hello,
When I want to "Configure Shadow Copies" on this Drive (Volume), the System tell me

"Volume xx is not eligible for shadow copies"

[HannesK]

that's what I expected to see... VAW relies on VSS snapshots, so it would work as soon as Microsoft supports VSS for DAX volumes

[GianlucaCroci]

For me (us) it's ok if we don't save this type of Drive. I don't mean that's not important, but TODAY it's not possible, and perhaps with a future version of Windows, or perhaps Veeam that will use another way to backup this memory, so we'll do it.
But TODAY what we can do? How we can backup a Server (Cluster) with this type of memories?

[HannesK]

today, you could exclude that drive from backup. The questions is what would happen with the logs on that drive

[GianlucaCroci]

Our problem is to exclude this type of Drive, because we backup ALL as a "Failover Cluster", we use as "Name Computer" the name of the Cluster, and the "Entire computer" as "Backup Mode".
So on the job we do'nt give all the disk that we want to backup, and I don't have seen where I can exclude Disks (with our Job configuration).
We've other jobs where we choose the Disks, but these servers aren't in a Cluster.
I don't know if I've explained myself

[HannesK]

Hello,
computer name of cluster is fine, but you could use volume level backup instead of entire computer



Best regards,
Hannes

[GianlucaCroci]

Thank's.
I know this, but it's a Cluster and a Disk that now is connected/associated to a Node, tomorrow can be connected/associated to another. What will do the Backup Job? It will return a Warning?

[HannesK]

the owner of the disk will be detected and the backup job "just works". As long as there is no issue with backup, there will be no warning.

[GianlucaCroci]

OK, thank's. I try and I'll tell you the result

[GianlucaCroci]

Hello,
Yesterday I've changed the Job.
Now the Job is configured with "Volume level backup" and I've specified all the Disk.

Tonight the job was executed and ended with many Warnings. In practice in the Job I inserted 10 disks, in addition to the disk C: which is common for all. In each server, Warning returned me on all the Disks that it didn't connect.
Being that in the Job I configured the saving of 2 different Clusters, they don't have the same Drive letters, and for this I've different Warning for each Node. The message is as follows:

Cannot backup volume H:\ because it does not exist

I've to redo the job and divide it. I think it's the only solution.

[GianlucaCroci]

GoodMorning,
with 2 job, one for each Cluster, it work good.
I'll try to add more DAX Volume, but I think that's the only solution at the moment.
Thank's a lot for the help.
Kind regards

[HannesK]

Hello,
good to hear, thanks for the feedback, as it could help others in future.

Best regards,
Hannes