How can I tell if a disk is failing in RAID?

Question

Iron

Windows 2003 system with terminal service.
Two WD screws are assembled in RAID-1 via Promise fasttrak tx2300
The other day, the system began to freeze, as I suppose, when accessing certain files. In general, everything works, but hangs a couple of times a day. Reboot helps.
Entries appear in Events in Windows:

Source: Ntfs
{Lazy Write Error} Unable to save all file data. Part of the data is lost. This error can be caused by a computer hardware or network connection failure. Try saving this file elsewhere.

Source: Application Popup Application Popup
: Windows - Lazy Write Error: Failed to save all data of file C:\******\****.*** Part of the data is lost. This error can be caused by a computer hardware or network device failure. Try saving this file elsewhere.

RAID status: Functional
SMART status of both screws: Healthy
So actually the question is, did the screws start to crumble or something with the RAM? Perhaps a raid controller?
And how can you test all this stuff?
Thank you.

Answer 1

we cut down the server, take out the screws, connect it to another computer, look smart and check the mhdd surface if everything is good with the screws, then
reboot the server, close access for everyone, run the file system test

Answer 2

It is not necessary to remove the disks from the server and check in another computer.
You can download the test system from a USB flash drive.
For example, SystemRescueCD with added MHDD and Victoria floppy disk images.

Answer 3

This controller is very similar to some kind of software. And judging by the xibt forum , it would be better to take out the disks, because. this line of RAID controllers does not provide SMART information, and only reports that the disk is dead or alive.
So bad sectors can be.