[Techtalk] Drive Failures

Anthony Gorecki anthony at futurepoint.com
Thu May 27 19:10:18 EST 2004


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

Over the past five months, I have had six separate hard drives fail either 
during the initial deployment of a server, or during its active operation. 
While two of the older drives failed for legitimate reasons, the remaining 
four failures are a cause for concern. 

In all cases, the drives would shut down at random times, some of which were 
shortly after powering on the system. All of the damaged drives were tested 
in normal systems to confirm that the failure was not local only to one 
server, and were ultimately unusable. 

Although data contained within the drives was left undamaged and could be 
accessed during the times where the drives were operational, they would 
continually power down, often for no identifiable reason. Operation of the 
drives would frequently resume for a period of a few minutes after an 
automatic reset from the kernel or BIOS, generally with mixed degrees of 
success, only to have the problem persist.

Having lost another drive this afternoon in the same system which housed the 
previous three failures, it's fairly obvious to assume that the problem must 
be a local one. Although I may have been willing to accept the previous 
failures of older drives to be an unfortunate string of bad luck, I don't see 
a viable explanation for the failure of a relatively newer drive, which has 
never had a history of reliability issues.

Other servers which are running on the same electical and internet lines have 
been functional for months with no abnormal hardware failures, therefore I 
believe building-wide dirty power is not a likely cause of the problem. 
Considering that no other hardware from the system in question has failed, I 
believe that the problem is local to the motherboard, rather than any of the 
other hardware.

Before the motherboard and power supply are replaced, does anyone have 
suggestions or a possible explanation for these hardware failures?


- -- 
Best Regards,
Anthony Gorecki
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.3.4 (GNU/Linux)

iD8DBQFAtp+N1YBSMGhoIMkRAtybAJ9PeajP0tzQiCqjTIZjOpNkFxTu2gCaAw5W
DjipEdGkhUYGYRlJ2bzUHbg=
=FycL
-----END PGP SIGNATURE-----


More information about the Techtalk mailing list