What happens when your NAS loses connection

At three in the morning, nobody is watching the recorder. So when the network drive it writes to quietly stops responding — a NAS reboots, a switch hiccups, a cable works loose — the failure that matters is not the drive going away. It is what your camera system does about it.

For a lot of recorders, the answer is: nothing good. And you find out days later, when you go looking for footage that was never there.

Why a dead drive takes the whole system down

Writing video to a network share looks simple until the share stops answering. When a networked disk goes silent, any part of the software that reaches for it does not get an error back — it just waits. And on many systems, that one waiting task is holding up everything else: the live view, the buttons, the part of the program that is supposed to notice something is wrong.

So the recorder does not crash. It does something worse. It freezes — live video included — while looking perfectly "on." No alert fires, because the code that would raise the alert is stuck in the same queue as everything else. By the time anyone notices, hours of would-be recordings are simply gone, and the first sign of trouble was a black screen when you needed the footage most.

Built to assume storage can fail

Calyston treats the recording disk as something that can and will misbehave — network share, USB drive, internal disk, all of it. Every operation that touches storage runs on a timer, off to the side, so that a disk which stops answering can never take the rest of the system with it.

That one design choice changes the whole failure story:

• The live view keeps working. Watching your cameras never depends on the recording disk being healthy. If storage stalls, video keeps decoding in your browser.
• You are told, in seconds. Instead of silence, you get a plain message — "storage is not responding — check the Storage page" — and, on paid tiers, a push notification to your phone. Within about half a minute of the drive going quiet, not days later.
• It recovers on its own. When the drive comes back, recording resumes automatically. No restart, and — just as importantly — no storm of false alarms the moment it returns.

That last point is subtle and it matters. A system that panics when storage comes back is almost as useless as one that ignores it leaving. Calyston knows the difference between "the disk just returned and is catching up" and "something is genuinely wrong," so a recovery is quiet, the way a recovery should be.

We did not just hope — we killed the drive

None of this is worth much as a promise. So we tested it the honest way: we set up a NAS, pointed a recording camera at it, and pulled the storage out from under a running system — three separate times, in different ways, while watching every internal signal.

Each round, the same result: the interface stayed responsive the entire time, the "check your storage" warning appeared within roughly half a minute, and when the drive came back the system picked up where it left off — with no false alarms on the way out or back in. The failures we found along the way got fixed and retested until the drive dying was, from the outside, a calm and clearly-labeled event instead of a silent black hole.

That is the whole idea. Your recorder should be at its most trustworthy exactly when something breaks — not just when everything is fine.

Get Calyston Community free →