Nimble Support

[L33T]

Hi,

Can I ask, does Veeam fully support Nimble storage arrays? I’d like some information on how they are integrated and if we can take advantage of snapshot backups to reduce our backup times?

Regards,
L33T

[Vitaliy S.]

Hello,

There is no integration with Nimble storage snapshots in v9, so you should be using regular backup/replication jobs to protect your VMs.

Thank you!

[L33T]

ok no worries.

[Vitaliy S.]

Just to update this topic - here is a good read on the upcoming integration > Advanced Integration with Nimble Storage coming in Veeam Availability Suite 9.5 for VMware!

[wumpus]

Hi Vitaliy,

Saw that support for VMWare will be available in 9.5 but will support be available for Hyper-V as well or in a later release? Thanks!

[veremin]

What kind of integration you're looking for? In case of Hyper-V we integrate with storage natively, at least with the ones having Hardware Provider VSS. (check Hyper-V subforum for more references)

Thanks.

[NoGoh]

Hi,

This seemed a suitable location for my comment on this topic. I am looking forward to the v9.5 release with the Nimble integration to make it easier for the client's backups then I can scrap my process below, yay. Hopefully it gives someone else pointers.

I've recently implemented v8 + patch3 (this was before v9GA was available) for a client running a pair of Microsoft 2012R2 Hyper-V Clusters each backed by their own Nimble Storage and using CSVs for VM storage. One Microsoft Hyper-V Cluster per site. v9 GA has not been implemented as of yet at the site.

Client wanted to use Nimble Storage snapshots and replication to replicate data between sites for DR purposes and to backup data from the remote site at the primary site. The link between the sites is a microware 300Mbps Link (a few Kilometers apart and they have numerous environmental issues and other congestion problems which effectively left me with not more than 25Mbps for replication on this link). A dedicated WAN link is not currently in the clients budget or physically possible nor was the Veeam WAN Accelerator feature available to me. Other factors limited what I could implement.

So the scenario is that VM data from the remote Hyper-V 2012R2 Cluster is Nimble Storage snapshotted and then replicated to the primary site using Nimble replication functionality. The trick was to backup this data at the primary site.

I was able to backup the VM data from the Nimble Replicated Snapshots and restore it successfully from both a Full and Incremental backup files.

Components Used:
-A Nimble Engineer provided me a starting point with some powershell scripts used for VMware based Nimble storage snapshots which I had to then "translate" into Hyper-V speak, still had to install VMware PowerCli for credential management else accessing the Nimble Array would all have been in clear text.
-A scheduled windows task to initiate the backup,
-Some additional coding to log all the powershell console output (and I do mean all output from the VEEAM job script, output from all the modified VMware/Hyper-V scripts, my master script, the remote Invoke-Command output). The logging also captures any errors where parts of the script may fail.
-The VEEAMultimatestartjobscript.ps1 by Andreas Neufert v1.06 (thank you so much for coding this, it saved me lots of time, and all I needed to do was to pass in the Veeam job name)
-VBR is a Virtual Machine running on the Primary site Hyper-V Cluster
-Repository is a physical server (was used for Backup Exec) backed by HP and Dlink SANs.
-Using all On-Host Hyper-V Proxies (customer had no budget for a dedicated off-host proxy and the Repository server was 2008R2)

Basically the process is as follows;
1. Remote Nimble performs a Storage Snapshot on the Hyper-V CSV volumes
2. Replicates said snapshot to Primary Site
3. Windows Scheduled task on the VBR VM kicks off once a night
4. Master script executes which,
4a. logs in to Primary Nimble Array,
4b. determines the second latest replicated snapshot from remote array because a replication may currently be taking place and it would be incomplete anyway
4c. Creates a Clone of that snapshot and provisions the iSCSI information on a designated Hyper-V host (which is part of the Primary Hyper-V Cluster)
4d. performs Microsoft Disk Management to bring that Cloned Volume online and register the VMs in the Hyper-V Host Manager using "Invoke-Command" from the VBR VM.
4e. Calls and executes VEEAMultimatestartjobscript.ps1 which inturn calls a pre-configured backup job in Veeam and waits for it to finish (using Forever Forward Incremental method).
4f. When finished it de-provisions the Remote VMs and Volume from the designated host
4g. Removes the Clone from Nimble storage and cleans up after itself
5. Master Script finishes all logging to a log file and quits.

What I found though is that I had to disable CBT for the Veeam Job which effectively forces Veeam to re-read the entire VM, but I still get an incremental vib for restoration purposes out of it but the backup can take hours. Not good, but I'm only backing up 5 VMs from the remote site anyway so time is not a problem, plus I could do it during the day if I had to at the primary site.

With CBT enabled in the job, the process would execute just fine, but come restoration, the data would only be available from the initial Full Backup (vbk) file, I could not select any data in subsequent vib files for restoration which was concerning (and the generated vib files were all the same file size). It seems to me that even though the Veeam backup is successful, the VM data from the snapshot clone volume is not read thoroughly when the volume is first mounted. Hopefully the v9.5 Release with Nimble Integration addresses this little foible but its something to be aware of.

Regards,

NoGoh