How many hard drive errors before you freak out?

Factual Questions
1

I’ve got multiple units of a system running XP Embedded that I maintain.
If I see more than a half-dozen error messages indicating failures to read/write during any one day, I replace the SATA drive the system uses.
The errors I usually see are along the lines of:
Windows System Log

Log Date Time / Jul 02 11:31:13 2008
Type: Error
Source: iaStar
Category: (0)
Event: 262153

Message:
The device, \Device\Ide\iaStor0, did not respond within the timeout period.

The rig makes occasional writes to an SQL database, and also writes about 130,000 JPEG files of about 16K apiece to an NTFS partition.
Aside from relying on SMART conclusions, are there any industry standards for how many messages like that a sysadmin should tolerate?

2

Anyone think I should have posted this in MPSIMS or something?
48 views, no responses…

3

How well are you backed up? With storage as cheap as it is, I would have zero tolerance.

Hard drives break a lot of hearts.

Could the hard drive be going to sleep and not waking up fast enough for Windows liking?

4

For certain logistical reasons, backups are not done on these rigs excepting the 2 MB config files.