Storage Capacity Planning: How Much to Buy and When
Storage fills faster than anyone plans for, and running out is disruptive in ways that are hard to fix quickly. This covers sizing storage so that does not happen.
Measure the growth rate, not the current size
The number that matters is how fast it is filling.
Current usage tells you where you are. Growth rate tells you when you run out, and that date is what a purchase has to be timed against.
Measure over long enough to be meaningful — several months at least, since growth is uneven. And watch for the pattern that catches people: growth accelerating rather than steady. A project, a new system or a retention change can double the rate without anyone noticing until the alert fires.
The overheads that consume capacity
Raw capacity is never what you get. Five things take a share.
RAID parity. Double parity across eight drives costs two drives of capacity.
Filesystem overhead, which varies by filesystem and block size.
Snapshots, which consume space proportional to how much changes and how long they are kept. This one grows quietly.
Reserved space, which some systems hold back and do not present.
The difference between TB and TiB. A drive sold as a number of terabytes presents fewer tebibytes, and across an array the difference is substantial. Our TB vs TiB guide covers why.
Together these mean usable capacity is considerably less than the sum of the drives, and plans built on raw figures come up short.
Do not fill storage
An important practical limit.
Most storage systems degrade as they approach full — fragmentation rises, writes slow, and some filesystems behave badly above a threshold. Copy-on-write systems in particular need free space to work.
So plan against a practical ceiling rather than total capacity, and treat approaching that ceiling as the trigger to act rather than a warning to note. Running at the limit is slower and riskier than running with headroom.
Plan the expansion path before you need it
Three questions to answer at purchase, not at the point of running out.
Are there free bays? If not, expansion means replacing drives or adding an enclosure. Our adding storage guide covers the options.
Can the array be expanded, and how long does that take? Expansion restripes data and leaves the array vulnerable throughout, which is a scheduled operation rather than an urgent one.
Will the drives still be available? Matching an existing array means the same capacity, class and sector format, and those get harder to source over time. Our availability guide covers why buying spares early matters.
Where the space actually goes
Before buying capacity, find out what is using it. Four things that are usually larger than expected.
Old backups and snapshots kept beyond their retention.
Logs with no rotation.
Duplicate data across shares and projects.
Data nobody owns — departed users, finished projects, systems decommissioned years ago.
Frequently a substantial share of a full array is reclaimable, and reclaiming it is cheaper than buying capacity. Our storage space guide covers finding it.
Buy for the horizon, not the moment
Three rules.
Size for the growth rate across the intended life, plus headroom, plus the overheads above.
Leave bays free where you can. Filling every bay at purchase means the next expansion discards drives.
Buy the capacity step that is best value, which is usually one or two below the largest available. Our 32TB guide covers why the top capacity carries a premium.
And remember rebuild time scales with drive size. Fewer larger drives is fewer bays and longer rebuilds; more smaller drives is the reverse. Our RAID guide covers the trade.
Frequently asked questions
How do I work out how much storage to buy?
From the growth rate across the intended life, plus headroom, plus the overheads that consume raw capacity — parity, filesystem, snapshots, reserved space and the TB to TiB difference. Current usage tells you where you are; growth tells you when you run out.
Why is my usable capacity so much less than the drives?
RAID parity, filesystem overhead, snapshots, reserved space, and the difference between terabytes as sold and tebibytes as presented. Together these mean usable capacity is considerably less than the sum of the drives.
Can I fill storage to capacity?
Not without consequences. Most storage degrades as it approaches full — fragmentation rises, writes slow, and copy-on-write systems in particular need free space to work. Plan against a practical ceiling rather than total capacity.
Should I check what is using the space before buying more?
Yes. Old backups and snapshots past retention, unrotated logs, duplicate data and data nobody owns are usually larger than expected. Frequently a substantial share of a full array is reclaimable, and that is cheaper than buying capacity.
Should I fill every drive bay at purchase?
Leave bays free where you can. Filling every bay means the next expansion requires discarding drives or adding an external enclosure, both of which cost more than planning for it.
Fewer large drives or more small ones?
Fewer larger drives uses fewer bays and less power but rebuilds slowly, exposing the array for longer. More smaller drives rebuild faster at the cost of bays and power. The capacity step one or two below the largest is usually the best value.
Tell us your current usage and growth rate and we will size storage against when you run out rather than what you have today.
