Dell PowerMax Operate (DCA Operate): Capacity Management and Reporting

Capacity questions test whether you report the correct number and can defend a growth decision.

Capacity vocabulary

• Usable: the capacity an SRP can actually serve. Free: what remains. Used: physically allocated to written tracks. Subscribed: the sum of thin device sizes presented into the pool.
• Subscription above 100% is normal for thin provisioning. The dangerous pattern is high subscription combined with fast used-capacity growth, not subscription alone.
• Effective capacity reflects data reduction (compression and, on PowerMaxOS 10 platforms, inline data reduction) and is not the same as usable physical capacity. Never compare an effective figure with a physical figure without stating the reduction assumption.
• SnapVX snapshots, linked targets and clones all consume SRP capacity, so protection overhead belongs in the report.

Where to look

• Unisphere capacity dashboards show SRP usable, used, free and subscribed, plus per-storage-group consumption and trend.
• SYMCLI options: symcfg list -sid 123 -srp for pool summary, symcfg -sid 123 list -capacity -thin for thin capacity figures, symdev show for a single device's allocated tracks, and symdisk list for physical drive state.
• Export or script reports through the REST API for regular capacity reporting rather than manual screenshots.

Worked example

An SRP reports 100 TB usable, 60 TB used and 130 TB subscribed. A team requests 30 TB more, which would take subscription to 160%. Used capacity will not jump at provisioning; it grows only as hosts write (plus snapshot change tracking). So the decision is not "no, oversubscribed". Check the used-capacity growth trend, protection overhead and alert thresholds: if growth is 3 TB per month, roughly 13 months of headroom remain before 100% used, so approve with capacity alerts at 80% and 90% and a plan to add capacity or migrate a workload. Deny only if the trend shows exhaustion before the workload can be moved.

Traps

• Reporting subscribed capacity as consumption, which overstates usage and triggers pointless projects.
• Omitting snapshots, clones and SRDF copies from the capacity picture.
• Mixing FBA (GB/TB) and CKD (cylinders/tracks) units in one report.
• Quoting effective capacity as if it were usable physical capacity.
• Waiting for "pool full" instead of alerting on used-capacity thresholds.