Troubleshoot disk drive failures

[rcllbrg]

I've had lots of disk drive problems for a while and I suspect the PSU(s)

Two 1year old Maxtors just died out of nowhere again, before it was WD drives dying progressively

Some drives were first on a Sparkle Power 350w which was supposed to be a decent generic PSU, then to an Enermax 420w

But I look at the Enermax voltages in the bios and they seem to be off, +12 is at 12.27 the +5v is at 4.93, ... if the +5v undervolts by that amount can it damage disk drives?

I heard you need electrical meter to really test the voltage accurately, but maybe there is a way to tell from the smart values on the dead drive what the failure was, does a particular kind of failure come from undervoltage? like heads crashing down on the platter because they got their power lowered?

I want to monitor these new drives as closely as possible, from the filedates I've had data corruption on the old drives well before they showed a smart error, Is there a recommended program to monitor hard disks for file corruption and failure prediction?

 

[imported_Phil]

Don't rely on the voltages displayed in the BIOS, they're never that accurate. If you want to measure them to check your PSU, get a cheap multimeter.

It sounds more like your drives are overheating. How warm are they to the touch? Do you have a fan cooling them?

Oh, and the heads won't crash down on the platter if the voltage lowers. The worst that will happen is that the motor in the drive switches off, and the heads automatically move to a "safe zone" on the disk- commonly referred to as a head park. This is what happens when the drive shuts off, and thus it's perfectly safe. Any data being written at the time may be corrupted though.

 

[rcllbrg]

I read something about degredation of a smart value for "head flying height", would lower voltage cause the head flying height to decrease and maybe collide or will it run fine until a >%5 undervolt happens and shuts the motor down entirely?

There's an 80mm fan in the front of the case but the drives have always been pretty warm to the touch, do they report their temperature somehow?

For monitoring drive integrity against corruption, does software make a checksum when the file is written and check against it constantly? wouldnt something like that overuse the disk and eat ram etc

I know NTFS itself has some sort of protection against corruption but I'd think it's a good idea to keep an eye on the drives specifically if there is recommended software for it

 

[rcllbrg]

I ended up installing the trials of both hmonitor and activesmart, they both report the same thing harddrive-wise

They show temperature and the rest of the smart attributes, it seems the drives max around 50c under heavy disk activity

So these are two good apps to poll the drive's smart attributes, but what can be checked for actual datafile corruption?

I can't seem to find a program for data integrity, besides Diskeeper which seems to optimize filewrite locations... but I don't think it has a way to monitor for file corruption