Forum Discussion

Allynz's avatar
Allynz
Everpure
8 hours ago

Fusion preset driven backup SLA

 

Using presets as the policy source of truth across vCenter and Veeam

 

Sharing a pattern I have been running in my lab: using an Everpure Fusion preset as the single policy source that drives storage provisioning, vCenter placement, and backup SLA in one motion.

The idea started from a simple frustration. We already encode a lot of intent into a Fusion preset: the array and pod, the protection group, snapshot schedule, replication, QoS. Then we throw most of that context away the moment the volume becomes a datastore, and a human re-decides protection downstream in a different tool. I wanted the preset to stay the contract all the way through, so provisioning once gets consistent storage and the right data protection without a second decision.

The flow:

  • I ask for a datastore of a given size from a named preset. Fusion deploys the volume with the preset’s protection group and settings.
  • Provisioning mounts it, formats VMFS, and exposes it in vSphere, with a pre-flight check across the cluster hosts and a placement recommendation pulled from the fleet before any write.
  • I assign the datastore a tier, which lands as a vCenter tag. That tag is what the backup vendor keys off (Veeam here), so the SLA follows the storage.

What stood out from a Pure angle:

  • The preset as source of truth. I can clone the protection group at provisioning time, change snapshot frequency or replication per workload, and the downstream tag and backup follow. The preset is not just a storage shortcut, it is the policy definition, and everything else derives from it.
  • Fleet-wide placement. Rather than guessing which array has room, the recommendation came from the fleet: the target array reported 4% used with host visibility confirmed up front. At fleet scale this is the difference between consistent tiering and a pile of one-off decisions that drift apart array to array.
  • Immutability that lines up across layers. SafeMode and Object Lock on the array pair naturally with an immutable backup repository on the other side, so the higher tiers get immutability at both the primary and the backup copy. For a ransomware conversation, being able to say tier 1 is immutable on the array and in the backup, by policy, not by hope, is the whole point.

The governance angle is what I think matters most. When the preset defines the tier and the tier is enforced automatically downstream, “tier 1” means the same thing on every array and for every workload. You stop relying on the operator to remember what tier 1 is supposed to look like this week.

I orchestrated it with MCP servers (Fusion, vCenter, Veeam) so every change is gated behind an approval, and it caught a real conflict mid-run, a protection-group name collision, and offered cleanup versus redeploy instead of failing. It even flagged the root cause (the protection group name is built from the first 25 characters of the workload name) so future builds avoid it. But the Fusion mechanics underneath are standard presets; the orchestration is just what made it conversational and safe.

Full write-up with architecture and screenshots: https://thebetasociety.com/wcith/pure-storage/everpure-mcp-policy-from-preset-to-veeam-backup/
Video walkthrough: https://youtu.be/FAVeVXPJECo

Would love to hear how others are using presets as a policy layer, especially anyone tying Fusion presets to downstream systems like backup, CMDB, or showback.

No RepliesBe the first to reply