HD failures and heat

[Wolfala]

I have a question.

We have a RAID node thats been operating for about 8 months now. During that time, 2 drives have gone down. 1 went down recently when a power outtage forced the server to restart. The questions are:

1.  Why would mearly restarting a box, regardless of the type fry a drive?

2.  When they quote MTBF of 50,000 hours, what is a normal expected life of a drive that is up 24/7 in a RAID array?

3.  If drives are stacked together, say 5 on top of another - can heat play a decisive factor in their lifespan and any other 2 problems mentioned?

Wolf

[Skuzzy]

Couple of things.

1)  If you are using IDE drives in a RAID, you should be aware none of the IDE/SATA drives on the market are rated for 100% duty cycle.

2)  Yes, absolutely, if you stack drives tightly together without insuring proper air flow between them, they will die sooner.  Heat is, without a doubt, the primary killer of a hard drive.

When you run drives, which are not rated for 100% duty cycles for long periods of time, and then turn them off, there is a fair chance they will never start again.  It is usually related to startup current requirements and how much the drive can deliver to the spindle motor (i.e. the drive spindle motor needs more current than the drive electronics can supply to it).