Disk failures can be limited, progressive or catastrophic. If the alerts are being received but the machine is still usable then it isn't catastrophic. Yet.
The first thing to do is run a full chkdsk. If there are bad sectors then this will mark those as such and avoid using them which may clear the alert. In this case the failure is limited and the rest of the drive may continue to operate ok for the remainder of its life. If the alert comes back or a chkdsk can't fix the errors then the failure is likely progressive so will get worse over time. There's no way to tell how long it will take to catastrophically fail if the failure is progressive, so it's just a matter of making sure a suitable backup is in place if there's anything on the device that isn't stored elsewhere. It's up to the client whether they replace sooner or wait till the device becomes unstable or fails completely.
Best to run the chkdsk at the end of the working day and leave it overnight, check the result in the morning - it can take some time to run. Arrange a suitable time with the client and then carry out the following:
- At a command prompt:
- chkdsk c: /f /r
Substitute c: for another device as appropriate. If asked to scan on reboot, say yes.