{{ slot.marker }}
Use installed or planned bays, e.g. 8 for an eight-bay NAS shelf.
Enter one drive's advertised raw size, then choose decimal TB/GB or binary TiB/GiB.
{{ manualDisks ? 'Fill each installed slot; leave empty bays blank so they are ignored.' : 'Every non-spare slot uses the shared disk size and selected unit.' }}
{{ manualDisks ? 'On' : 'Off' }}
Enter capacities in the selected unit; blanks are empty bays, and the smallest engaged disk caps usable space.
{{ manualPreview.installedCount }} installed {{ manualPreview.emptySlots }} empty slot(s) Smallest {{ formatCap(manualPreview.smallest) }} {{ unit }} Largest {{ formatCap(manualPreview.largest) }} {{ unit }}
Supported: RAID 0, 1, 5, 6, 10, 50, and 60; nested choices reveal extra controls.
Use whole numbers: RAID 1/10 need mirror width, RAID 50/60 need group count and disks per group.
Mirror width
Mirror width
Use 0 for no dedicated spare; mixed arrays reserve the largest spare candidates first.
drives
Presets set the overhead slider to 2% ext4/XFS, 5% btrfs, or 8% ZFS.
{{ formatPercent(fsOverhead, 1) }}
Set 0% for raw RAID math; use Custom profile when entering your own percentage.
{{ formatPercent(targetFill, 0) }}
100% uses all calculated usable space; 80-90% leaves room for snapshots and rebuild work.
{{ formatPercent(safetyReservePercent, 0) }}
Default 5%; raise it when the number will be promised to applications or teams.
Leave 0 unless each active disk has a known reserved partition; choose GB/GiB or TB/TiB.
Use a sustained MB/s estimate, e.g. 120-220 for many HDD rebuilds.
MB/s
Enter AFR as percent, e.g. 1.5 for 1.5% per disk per year.
%
Metric Value Copy
{{ row.label }} {{ row.value }}
Metric Value Copy
{{ row.label }} {{ row.value }}
Layout Planned usable ({{ unit }}) Tolerance Idle disks Rebuild Risk Verdict Copy
{{ row.raidLabel }}
{{ row.note }}
 {{ row.displayPlannedUsable }} {{ row.toleranceDisplay }} {{ row.idleDisplay }} {{ row.rebuildDisplay }} {{ row.riskDisplay }} {{ row.statusText }} 

                

            


        

    

                

                

                    

                        

                            Tags: Storage, Sysadmin
                        

                        

                            
                            
                        

                    

                

            

        

    


                    
        

        

        

            

            
        

        


                    

                

                

                    

                                                                                

                                
                                    
                                        RAID Planning Suite
                                                                                    Plan a RAID layout with usable capacity, workload headroom, rebuild and URE risk, growth runway, charts, and downloadable reports.
                                                                            
                                    
                                
                            

                                                                                

                                
                                    
                                        RAID Performance Calculator
                                                                                    Estimate RAID safe IOPS and throughput from drive benchmarks, write mix, spares, controller caps, degraded behavior, and target margin.
                                                                            
                                    
                                
                            

                                                                                

                                
                                    
                                        RAID Reliability and MTTDL Calculator
                                                                                    Estimate RAID MTTDL from AFR, rebuild time, URE rate, spares, and fleet size with loss probabilities, mitigation options, and tradeoff charts.
                                                                            
                                    
                                
                            

                                                                                

                                
                                    
                                        RAID Sizing Calculator
                                                                                    Calculate RAID and RAID-Z usable capacity with spares, fill headroom, chassis fit checks, drive-size ladders, charts, and rebuild risk estimates.
                                                                            
                                    
                                
                            

                                            

                

            

        
                                                                                                                                                                                                                                                        
        

        

            
            

                Customize

                

                    

                        

                            
                            
                        


                        

                            
                            
                        


                        

                            
                            
                        

                    


                    

                        Advanced

                        

                            

                                
                                

                                    
                                    
                                

                            


                            

                                
                                

                                    
                                    
                                

                            


                            

                                
                                

                                    
                                    :
                                    
                                

                            


                            

                                
                                

                                    
                                    
                                

                            


                            

                                

                                    
                                    
                                

                            


                            

                                

                                    
                                    
                                

                            


                            

                                
                                
                            


                            

                                
                                
                            

                        

                    

                

            

            
        


        

            
            
            
        


        
        

    

    

        

  
A storage shelf rarely turns every advertised terabyte into space that applications can safely fill. Redundant Array of Independent Disks (RAID) layouts trade raw capacity for striping, mirrors, parity, spares, and recovery behavior. The headline size on a drive label is only the starting number; the useful planning figure is what remains after the array geometry and operating headroom have done their work.


  
Capacity planning becomes harder when the array has mixed drive sizes, a hot spare, or a nested layout such as RAID 10, RAID 50, or RAID 60. A larger disk can be partly wasted when it joins a protected group with smaller disks. A spare can be valuable for faster recovery, but it must not be counted as usable space. A parity layout can fit more data than a mirror-heavy layout, but wide single-parity groups spend more time exposed during rebuilds.


  

    

      
      
Planning questionWhat changes the answerWhy it matters

      
        
How much space can be used?Disk count, smallest engaged disk, RAID level, spare count, and reserved headroom.Raw installed capacity can be much larger than the capacity that should be assigned to shares, databases, or virtual machines.

        
How much protection is guaranteed?Mirror width, parity count, group count, and whether failures land in the same group.Best-case tolerance can sound safer than the guaranteed failure count that an operator should plan around.

        
How stressful is recovery?Drive size, rebuild throughput, annual disk failure rate, and the number of disks exposed during rebuild.Larger drives and slower rebuilds keep the array in a degraded state for longer.

      
    

  


  
RAID terms often hide two different ideas. Usable capacity is the amount available after redundancy math. Planned usable capacity is more conservative because it also leaves room for filesystem metadata, snapshots, growth, and emergency reserve. Those extra reductions are not wasted space; they prevent a storage plan from being filled to the point where rebuilds, snapshots, or normal growth become risky.


  

    RAID capacity sketch showing raw disks reduced by spares, protection, overhead, fill headroom, and reserve before planned usable space remains.
  


  
The safest RAID capacity answer is still a planning estimate, not permission to skip backups or platform checks. A controller, NAS appliance, filesystem, or cloud storage gateway may impose its own group limits, metadata overhead, rebuild behavior, and drive qualification rules.






  
How to Use This Tool:

  
Enter the disk inventory first, then compare the selected RAID layout against the capacity ledger, failure envelope, and matchup views.

  
    
    
  1. Set Disk slots, Disk size, and the unit. Use decimal TB/GB for drive-label planning or binary TiB/GiB when your source figures already use binary units.
    2. 
    
  2. Turn on Define each disk manually when the shelf mixes capacities. Blank manual slots are treated as empty bays, and the preview shows installed count, empty slots, smallest disk, and largest disk.
    3. 
    
  3. Choose RAID layout. For RAID 1 and RAID 10, set mirror width. For RAID 50 and RAID 60, set group count and disks per group.
    4. 
    
  4. Add Hot spares before reading capacity. Manual mixed-disk plans reserve the largest spare candidates first so standby drives do not inflate usable space.
    5. 
    
  5. Open Advanced when you need filesystem overhead, target fill, safety reserve, per-disk system reserve, rebuild throughput, or annual disk failure rate to match your planning policy.
    6. 
    
  6. If a red warning says the layout is not buildable, fix the named problem before trusting the selected layout. Common causes include too few active disks after spares, every installed disk being assigned as a spare, or RAID 50/60 group geometry that needs more active disks.
    7. 
    
  7. Use the topology strip to confirm data, parity, mirror, spare, and idle roles, then read Capacity Ledger, Failure Envelope, RAID Matchup Chart, RAID Matchup Table, Capacity Mix Chart, and JSON for the details you need to keep or compare.
    8. 
  

  
For a purchase or migration decision, save the current assumptions along with the result. RAID capacity numbers are only comparable when units, spares, overhead, fill target, reserve, rebuild throughput, and failure-rate assumptions stay the same.






  
Interpreting Results:

  
Start with Planned usable after reserve, not installed raw capacity. That field already accounts for the selected RAID layout, overhead, target-fill headroom, and safety reserve. Then check whether the layout posture, failure tolerance, idle disks, and rebuild estimate agree with the risk you can accept.

  

    

      
      
Result cueMeaningWhat to verify

      
        
Planned usable after reserveThe conservative capacity figure after redundancy, overhead, target fill, and reserve are applied.Use this for commitments to users or workloads.

        
Mixed-size lossEngaged disks are equalized to the smallest engaged capacity, so larger members can strand space.Check whether regrouping disks or standardizing sizes recovers enough space to matter.

        
Idle active disksActive disks remain outside the final geometry because mirror width, group count, or disks per group does not divide cleanly.Adjust layout detail or spare count before ordering hardware.

        
Guaranteed disk failures toleratedThe conservative failed-disk count the layout can survive without relying on failures landing in favorable groups.Treat this as the operational acceptance number.

        
Best-case disk failures toleratedExtra failures may be survivable only when they occur in different mirror or parity groups.Do not use best-case tolerance as a backup or recovery promise.

        
Long rebuild windowThe posture appears when rebuild time reaches 36 hours or the conservative rebuild-window risk reaches 0.5%.Compare RAID 6, RAID 10, smaller groups, faster media, or a hot spare.

      
    

  

  
A high capacity result can still be the wrong plan. If the selected layout has no redundancy, a wide single-parity group, many idle disks, or a rebuild longer than a day, use the matchup table and failure envelope before accepting the space gain.






  
Technical Details:

  
RAID capacity math starts by deciding which disks are actually engaged. Hot spares are removed first. Blank manual slots are ignored. Nested RAID 50 and RAID 60 layouts engage only the full groups described by group count and disks per group, so extra active disks can remain idle when the geometry does not divide evenly.


  
Protected arrays normally size each participating member from the smallest engaged disk. That equalized member size prevents a larger disk from increasing protected capacity inside the same group. After equalization, the model subtracts per-disk system reserve and filesystem overhead, applies the RAID layout efficiency, then applies target fill and safety reserve to produce planned usable capacity.


  
Formula Core:

  

    
      
        Ceq=Nengaged×Smin
        Cpost=(Ceq-Rsystem)×(1-Ofs)
        Cprotected=Cpost×Elayout
        Cplanned=Cprotected×Ftarget×(1-Rsafety)
        Trebuild=SminVrebuild
        λ=Nexposed×AFR×Trebuild8760
      
    
  


  

    

      
      
LayoutCapacity efficiency ruleGuaranteed tolerance rule

      
        
RAID 0All equalized engaged capacity is usable before overhead and reserves.0 disks.

        
RAID 11 / mirror width across complete mirror sets.Mirror width minus 1.

        
RAID 5(N - 1) / N for one distributed parity member.1 disk.

        
RAID 6(N - 2) / N for two distributed parity members.2 disks.

        
RAID 101 / mirror width across striped mirror sets.Mirror width minus 1.

        
RAID 50(disks per group - 1) / disks per group across complete RAID 5 groups.1 disk conservatively; more only if failures land in different groups.

        
RAID 60(disks per group - 2) / disks per group across complete RAID 6 groups.2 disks conservatively; more only if failures land in different groups.

      
    

  


  

    

      
      
SymbolMeaningSource in the calculation

      
        
NengagedDisks participating in the selected layout after spares and idle disks are excluded.Disk inventory plus RAID geometry.

        
SminSmallest engaged disk capacity in the selected unit after conversion to bytes.Shared disk size or smallest populated manual disk in the engaged set.

        
RsystemTotal per-disk system reserve removed from equalized engaged capacity.System reserve per active disk multiplied by engaged disks, capped at equalized capacity.

        
OfsFilesystem overhead fraction.Custom value or the ext4/XFS, btrfs, or ZFS preset.

        
FlayoutRAID layout efficiency.Mirror, parity, or nested group rule.

        
AFRAnnual disk failure rate as a decimal fraction.User-entered annual disk failure rate divided by 100.

      
    

  


  
The rebuild-window risk is a conservative comparison signal based on annual failure rate, estimated rebuild duration, and the number of exposed disks. It is not a full reliability forecast because it does not model drive age, workload, controller behavior, latent sector errors, scrubbing, correlated failures, or vendor-specific rebuild algorithms.






  
Limitations:

  
RAID capacity estimates are useful for comparing layouts, but they do not certify that a storage platform can build or recover the exact array. Before using a result for a production plan, confirm these items:

  
    
    
  • Controller, NAS, filesystem, or operating system support for the selected RAID level and group geometry.
    • 
    
  • Drive qualification, sector format, firmware behavior, and sustained rebuild throughput for the actual hardware.
    • 
    
  • Backup and restore policy, because RAID protects availability for some disk failures but does not replace backup copies.
    • 
    
  • Performance needs, since capacity math does not estimate write penalty, latency, cache behavior, or workload contention.
    • 
  






  
Worked Examples:

  
Eight 12 TB disks in RAID 6 with no hot spare start with 96 TB installed raw capacity. RAID 6 reserves two disks worth of parity, so protected capacity is about 72 TB before overhead. With 5% filesystem overhead, an 80% target fill, and a 10% safety reserve, Planned usable after reserve is about 49.25 TB.


  
The same eight 12 TB disks with one hot spare engage seven disks instead of eight. RAID 6 then has about 60 TB protected capacity before overhead, and the same 5% overhead, 80% target fill, and 10% reserve produce about 41.04 TB Planned usable after reserve. The spare lowers usable capacity, but it can let recovery begin sooner after a failed disk is replaced logically.


  
A wide RAID 5 set can look attractive on space. Eight 12 TB disks in RAID 5, using 180 MB/s rebuild throughput and the default 5% safety reserve with no filesystem overhead, yields about 79.80 TB Planned usable after reserve. Reducing rebuild throughput to 80 MB/s pushes Estimated rebuild duration above 36 hours, which changes Planning posture to Long rebuild window even though the capacity number is still high.


  
A RAID 60 plan with two groups of four disks needs eight active disks. If an eight-slot plan also assigns one hot spare, the selected layout reports that it needs eight active disks for two groups of four. Remove the spare, add another installed disk, or choose a layout such as RAID 6 that fits seven active disks before reading Capacity Ledger for the selected layout.






  
FAQ:

  
Why is usable capacity lower than the drive labels add up to?
The calculation removes hot spares, idle disks, mixed-size loss, RAID redundancy, filesystem overhead, target-fill headroom, and safety reserve before reporting Planned usable after reserve.

  
Should I use TB or TiB?
Use TB, GB, MB, or PB for decimal drive-label planning. Use TiB, GiB, MiB, or PiB when your source numbers already come from binary operating-system or storage-platform reports.

  
Why does a larger disk not always add more usable capacity?
Protected RAID groups are sized from the smallest engaged disk in the set. The larger disk's extra capacity appears as Mixed-size loss unless the grouping or disk set changes.

  
Why does the selected layout say it is not buildable?
The warning text names the failing condition, such as too few active disks after spare allocation, all disks assigned as spares, or RAID 50/60 groups that need more active disks than the plan has.

  
Does RAID replace backup?
No. The Failure Envelope describes disk-failure tolerance for the selected layout, not protection from deletion, ransomware, controller faults, filesystem corruption, site loss, or application mistakes.






  
Glossary:

  

    
Engaged disk
A disk that participates in the selected RAID geometry after spares and idle active disks are excluded.

    
Equalized capacity
The engaged disk count multiplied by the smallest engaged disk capacity.

    
Hot spare
A standby disk held out of normal capacity so it can be used when a member fails.

    
Planned usable capacity
The conservative capacity left after RAID redundancy, overhead, target fill, and safety reserve.

    
Rebuild window
The estimated time needed to rebuild a failed member from the smallest engaged disk size and rebuild throughput.

    
Guaranteed tolerance
The failed-disk count the selected layout can survive without relying on failures landing in favorable groups.

  






  
References: