Repository Feature Request

[dkvello]

We have re-designed (new HW and more) for our V7 Ent installation. We have set up two new proxy's (W2K12 Std) with plenty of CPU and Memory, a bunch of 8GB FC HBA's and 10GB Eth NIC's.

Primary vSphere Storage is IBM SVC/Storwize V7000 Fiber Channel and Target is Nexsan e48 Fiber Channel. So the proxy's read from FC storage and writes to FC storage.
The W2K12 nodes are set up to use CSV (Clustered Shared Volumes) disk from the Nexsan. This gives us a unified file-system without using the network/SMB.

Now, the backup jobs can be set up to pick Proxy based on load/availibility (Auto), but jobs are bound to repositories and in this case is then bound to the proxy that owns the repository.

I'd like the repository service to be cluster aware in situations like this. i.e. I want a feature that would have a repository definition span all the proxies in a W2K12 cluster with CSV storage.

This way all proxies have access to the same repositories and a job can pick a clustered proxy based on load and not on what proxy owns each repository. The repository ownership can then switch to another proxy if needed.

I know You can set up a separate SMB3 host as repository to achieve this, but that would mean having to move data across the network unnecessarily as well as introducing a bottleneck.

[Vitaliy S.]

I have heard a couple of the same requests/design requests already. The only way to achieve that now is to use network stack, but it is an interesting idea, indeed.