Include resynchronization time in S2D per-server capacity planning

Source facts

Microsoft recommends limiting the total storage capacity on an individual server to about 400 terabytes. The source connects greater capacity per server with longer data-resynchronization time after an outage or reboot. It lists a four-petabyte maximum storage-pool size, with a one-petabyte limit for Windows Server 2016. Microsoft Learn.

Applicability

Identify the server version and distinguish capacity on one server from capacity across the whole pool. Review the current platform limits before applying the numbers. Ask the workload owner what recovery and maintenance duration is acceptable for the proposed expansion.

DSE recommendation

Prepare a per-node capacity inventory and a separate pool total. Pair the expansion request with an observed or tested resynchronization-time estimate appropriate to the actual hardware and workload. Have the storage owner explain how the larger node affects the maintenance plan. Avoid presenting a supported maximum as the preferred design target without considering operational recovery.

Verification

Measure a representative approved node-return and resynchronization exercise before accepting the expanded design. Record data volume, workload, hardware, elapsed time, and health at completion. Compare the result with the maintenance and recovery expectations. Retain the measured conditions so future capacity additions trigger another review rather than inheriting an unrelated timing assumption.

Official references

Microsoft Learn: Choose drives for Azure Local and Windows Server clusters. Source reviewed September 8, 2026.