RAID selection is a business decision expressed through a storage configuration. It influences how installed capacity is used, how the system responds to specified drive failures and what operating conditions administrators need to monitor. The Huawei OceanStor Dorado 2020 lists RAID 5, RAID 6, RAID 10 and RAID-TP among its supported options. Choosing between them requires more than ranking the names by the amount of space they appear to preserve.
For a system supplied with twelve 3.84 TB SAS SSDs, begin with the raw total of 46.08 TB and a clear workload description. The final deployable capacity depends on the actual supported pool design, reservations and services. This guide explains the concepts behind the listed RAID options and the questions to ask ICTECH Distribution when requesting a configuration. It does not substitute generic RAID arithmetic for a model-specific sizing report.
Define the problem RAID is intended to solve
RAID combines storage resources in a layout designed to tolerate particular drive failures while continuing to provide access to data. Different layouts achieve redundancy through parity, mirroring or platform-specific implementations. The protection level needs to be understood together with the size and arrangement of the protected resources. A claim about failure tolerance should always state the conditions under which it applies.
Drive-failure protection is only one part of data protection. A valid write that overwrites the wrong file can be faithfully stored across a redundant layout. RAID does not identify the business intent of that write or automatically restore an earlier version. Plan point-in-time recovery and independent backups alongside the pool design so the organisation has procedures for both hardware incidents and unwanted data changes.
RAID 5 and single-parity concepts
In a conventional RAID 5 explanation, parity information allows a protected group to tolerate one drive failure. The arrangement provides a capacity-efficient introduction to redundancy, but its suitability depends on workload, recovery requirements and the supported implementation. After a failure, the system must maintain access and restore protection using its defined recovery process. That degraded operating period deserves attention in the design review.
Ask what happens to performance and risk while protection is being restored. Identify alerting requirements, replacement procedures and the resources used during recovery. A layout that looks efficient during healthy operation may be less attractive when the business considers the consequences of a second relevant failure or a prolonged degraded state. Base the decision on the complete operating scenario rather than only a healthy-state capacity estimate.
RAID 6 and dual-parity concepts
Conventional RAID 6 uses two parity equivalents and is designed to tolerate two drive failures within its protected group. That additional redundancy typically reduces the capacity available for independent data compared with an otherwise equivalent single-parity arrangement. For workloads with demanding availability requirements, the extra tolerance can be part of a broader resilience discussion, subject to the platform's actual layout and supported settings.
An arithmetic example with twelve equal 3.84 TB drives in one traditional RAID 6 group leaves ten drive equivalents, or 38.40 TB, before other deductions. The example explains parity cost; it is not a guaranteed Dorado 2020 usable figure. Obtain the actual configuration output and ask which resources are reserved for spare capacity, metadata and the services planned for the system.
RAID 10 and mirrored data
RAID 10 is commonly explained as striping across mirrored pairs. In that simplified arrangement, half the raw capacity is available before other overhead because data is mirrored. Twelve equal drives would therefore represent six data-drive equivalents, or 23.04 TB, in the conventional illustration. Actual support and configuration on the array must still be confirmed before using that figure in a purchase decision.
Failure tolerance in mirrored layouts depends on which drives fail. Losing one drive from several different pairs differs from losing both members of one pair. Avoid describing RAID 10 as tolerating an arbitrary number of failures without that placement qualification. Review the platform implementation, supported layout and workload behaviour with the configuration team, then compare the result against the business requirement and available capacity budget.
RAID-TP and three-drive tolerance
Huawei lists RAID-TP for the Dorado 2020 with tolerance of three simultaneous SSD failures under supported configurations. The supplied planning notes also call for checking minimum drive requirements, including an eight-SSD threshold for the stated arrangement. Verify that requirement against the exact release and proposed pool layout rather than treating a short product description as complete implementation guidance.
Triple-parity terminology does not justify assuming a universal raw-minus-three capacity formula for every array configuration. Pool architecture, supported stripe choices and reservations affect the outcome. Ask for a written protection and capacity statement for the twelve-drive configuration. It should describe the intended tolerance, any population constraints and the resources required to recover protection after a failure. This gives both procurement and operations a concrete design to review.
Consider workload behaviour alongside capacity
Applications differ in request size, concurrency, read and write mix and sensitivity to latency. Parity and mirrored protection arrangements can handle these demands differently depending on the implementation and system resources. Generic claims that one RAID level is always fastest can mislead when they ignore cache behaviour, drive count, pool design and the workload being measured.
Use representative application tests or supported sizing tools to compare the proposed options. Include busy periods and concurrent operations that occur in normal service. A database, backup repository and virtualisation cluster may not place the same pressure on the array. Keep the test conditions with the results so future administrators can understand whether a performance change reflects workload growth, a configuration change or a developing fault.
Review recovery behaviour before a drive fails
Protection recovery is an operational process with capacity, performance and staffing implications. Ask how the system detects a failed drive, how data is reconstructed and which conditions affect the duration of that work. Do not accept a fixed rebuild-time promise without a description of the conditions used to produce it. Workload activity, available resources and the specific fault can change the observed result.
Prepare replacement procedures and verify who receives alerts. Know which support entitlement applies, where replacement parts come from and who can install them. A resilient layout can still spend too long in a degraded state if nobody owns the response. Record the steps for confirming that the array has returned to a healthy and protected condition after component replacement or recovery activity.
Keep snapshots and backups in the capacity discussion
The selected RAID layout determines part of the capacity baseline. Recovery points, cloned datasets and temporary test copies add other demands. If HyperCDP or HyperSnap is part of the intended deployment, include retained changed data in the sizing exercise. Avoid presenting all protected usable capacity as available to production applications while also expecting recovery services to use that same space without a budget.
Independent backups should also be designed according to the business recovery requirements. They can provide another route when the primary storage environment is unavailable or affected by an incident beyond individual drive failure. Test that route separately. A comprehensive plan explains which mechanism addresses each scenario and how operators decide which recovery procedure to follow when an actual incident occurs.
Ask whether the design can change later
Storage requirements evolve, but changing a protection layout may have constraints or require a migration. Before purchase, ask what supported options exist for adding drives, expanding pools or moving data to a different arrangement. Include temporary capacity and change windows in the discussion. Do not assume a configuration can be converted freely just because the platform supports several RAID names.
Document the initial design and the likely next expansion step. If growth depends on specific drive increments or compatible enclosures, identify those parts in advance. This allows the organisation to compare lifecycle costs and operational effort, rather than selecting a layout that only looks attractive at the initial drive count. Future flexibility should be verified as a supported procedure with known prerequisites.
Make the final RAID choice reviewable
Provide ICTECH Distribution with the application mix, required usable space, growth forecast and consequences of interruption. Request a proposed Dorado 2020 protection layout with explicit assumptions. Ask for the resulting capacity, supported failure scenarios and operational responsibilities to be documented together. This gives decision makers a clear basis for comparing protection options without relying on broad labels or unsupported guarantees.
After installation, retain the accepted configuration and verify it against the management interface. Confirm monitoring, replacement procedures and recovery tests during handover. RAID selection is successful when the deployed system meets the intended protection design and the team knows how to operate it. Capacity efficiency matters, but it belongs within that larger and more practical definition of a suitable enterprise storage configuration.
Frequently asked questions
Which RAID options are listed for Dorado 2020?
The model lists RAID 5, RAID 6, RAID 10 and RAID-TP. Confirm availability and supported layouts for the exact system release and drive configuration.
Does RAID-TP mean any three failures are always safe?
Huawei describes three-SSD failure tolerance under supported configurations. Verify the actual protection boundaries, drive requirements and layout before applying that statement to a deployed system.
Are generic RAID calculators suitable for final sizing?
They explain basic trade-offs, but they may omit array-specific layouts and reservations. Use the supplier's configuration report for a final usable-capacity commitment.
Does RAID replace HyperCDP or backups?
No. Drive redundancy, earlier recovery states and independently maintained copies address different incident scenarios. Plan and test the mechanisms required by the business.
Which RAID level should my organisation choose?
Compare capacity needs, workload behaviour, tolerated failures and recovery operations. Request a supported design based on those requirements rather than choosing solely from a generic ranking.
Product details and enquiries
Huawei OceanStor Dorado 2020 storage system | LIC-D2020-CDP-S HyperCDP licence | D2020-SAS3.84T-CF enterprise SSD
Contact ICTECH Distribution with your capacity, host connectivity, compatibility and recovery requirements for a configuration-specific enquiry.
Configuration reference: supplied ICTECH Distribution product information and bill-of-material notes. Exact inclusions are subject to the final quotation and compatibility review.
