Harddisks Failing

[jescolta]

Here is my pc specs

q9300
asus p5q
antec neo he-380
msi hd3870

i have 5 hard-disks
2x samsung hd103sj 11 months and failed smart reports read errors and 500 pending sectors

1 seagate st3500418AS -1 year 173 relocated sectors count

1 seagate st31000524AS -2 months failed -clicking and very slow

1 western digital wd5000AAKX - 1 month failed - too many read errors
computer would hang while trying to copy any file from it


the harddrives run 24/7 with 2 65 cfm 120mm fan cooling them

i dont understand why these drives always fail
temperatures range from 24 degrees (during morning) to 38 degrees
anyone can help me troubleshoot the problem the drives are now sent to supplier for rma but im afraid when i use them again they will fail with the same problems

drives are running in ahci mode

using windows 7 x64 with complete updates

 

[VirtualLarry]

what power supply?

 

[jescolta]

antec neo he 380 watts
i know its old but i think it can provide enough power for the system

 

[code65536]

Clicking and slowness is often a mechanical issue.

But the rest of your problems are just bad sectors. Which do not necessarily indicate a physical fault. (It might be a physical fault, but that is not necessarily so.)

Consider this scenario. As your hard drive's head is hovering over a platter, there is a surge of power to the drive. This power surge can produce a significant flux of the EM field around the head. This causes random garbage bits to be inadvertently written to the area of the platter that the head happened to be hovering over when this happens.

Now, the next time you try to read these sectors, because random garbage data had been written to them, the checksums will not match, and you will get a CRC mismatch error (better known as a read error). Remember, all that the drive knows is if the checksum is correct or not. It doesn't have a camera to look at the platter and determine if the unexpected bits is the result of physical damage to the platter or the result of random data having been deposited there. The only way the drive knows if a read was successful is if the checksum matches; this is how it's possible to get read errors even if the drive is actually healthy and has zero physical defects.

If the drive later tries to write to the sector, then it has an opportunity to experiment with the sector and see if it really is a physically bad sector or if this was just a case of stray bits having been written. In the former case, the sector pending reallocation is reallocated. In the latter case, the sector is simply removed from the pending list without being reallocated.

So. If you do a low-level format of the drive to force a write to those sectors, do the pending-reallocation sectors become reallocated sectors or do they just disappear? If it's the latter case, and if you see this happening on more than one drive, then chances are, your drives are getting dirty power.

Dirty power to the drives might be the result of a lousy power supply or general dirtiness in the power in your area (a UPS might help here; also, a good PSU should be able to cope with some dirtiness in the mains power). The quality of the power provided by a PSU is just as important as the quantity.