You try to open a drive, copy a folder, or run a backup and Windows stops you with: “The request could not be performed because of an I/O device error.” Sometimes it comes with the code 0x8007045D.
I/O stands for input/output. What Windows is telling you is that it asked the hardware to read or write something and the hardware failed to do it. Unlike most Windows errors, this one isn’t about your files or your file system — it’s a report from further down the stack.
What it usually means
Bad sectors. The most common cause. The drive attempted the read, retried internally, and reported failure. That’s a physical media problem, and it’s closely related to the CRC error you may also be seeing — same underlying condition, different message depending on where the failure surfaced.
A failing head or degrading surface. If the error appears in specific regions rather than randomly, one head may be weakening.
Connection problems. A marginal cable, a failing USB bridge, or an underpowered port can produce genuine I/O failures on a perfectly healthy drive. Worth ruling out because it’s free.
Drive dropping off the bus mid-operation. If the drive is also disconnecting and reconnecting, the I/O error may be a symptom of that rather than a media problem.
The free checks, in order
Do these before assuming the worst. None of them write to the drive.
- Different cable. USB cables fail internally more often than anyone expects.
- Different port, directly on the machine. No hubs, no docks, no monitor ports.
- Different computer. Rules out a driver or controller issue on yours.
- Its own power brick, if it has one. Desktop externals brown out when the supply weakens, and the errors look identical to media failure.
If the error follows the drive across all of those, it’s the drive.
Free evaluation. Firm quote before any work begins.
What not to do
Don’t keep retrying. Every retry sends the head back to the same failing region for another round of internal retries. This is a slow way of making the problem permanent.
Don’t run chkdsk /r. The /r flag forces a read attempt on every sector on the disk. On a drive already returning I/O errors, that’s hours of maximum stress on hardware that’s failing right now, and I’ve seen it finish the job. Full explanation here.
Don’t run a S.M.A.R.T. long self-test hoping for information. Same problem — it’s a full-surface read on a drive that can’t do full-surface reads.
Don’t defragment. It reads and rewrites nearly everything on the volume. On a degrading drive it’s the worst possible use of its remaining life.
What to do instead
Power it down and image it.
The right tool for this reads in large blocks, skips past problem regions instead of grinding on them, and returns for the difficult sectors afterward with the healthy majority already captured. That ordering is why an imaging pass gets more off a failing drive than a copy operation ever will — a normal file copy stalls on the first bad sector and keeps stressing the drive while everything past it stays uncopied.
Once the image exists, the unreadable areas are a bounded, known quantity. Usually it’s a handful of damaged files out of thousands.
The bottom line
An I/O device error is the hardware saying it couldn’t complete the operation. Rule out the cable and the port, because that’s free. If the error follows the drive, stop retrying, don’t let any scanning tool loose on it, and get it imaged before the failing region grows.
Getting I/O errors on a drive that matters? Start a case — mention whether it happens on the same files each time or moves around. That one detail separates a localized bad patch from something broader.