xfssyncd and disk spin down

[Petre Rodan]

Hello,

On Thu, Dec 23, 2010 at 01:29:14PM -0600, Stan Hoeppner wrote:
> Petre Rodan put forth on 12/23/2010 10:55 AM:
> 
> > this hard drive has exceeded it's 300k load/unload maximum from the specs in only 140 days, which means it was woken up every 30s or so. not willingly.
> 
> Contact WD and request a warranty replacement for the drive due to
> exceeding the spec'd cycle max in less than 6 months.  They may not
> honor your request.  Whether they do or not, I'd replace the drive as
> the mechanism obviously has too much wear on it already, and within one
> year of use will have well over double the spec'd cycles.  If you
> replace it with another 20EARS, replace the firmware immediately as
> mentioned below to decrease the load/unload rate.  (It would be nice if
> they offered the ability to disable the sleep mode totally, but then it
> wouldn't be "green" pfft).

thanks for your input. I did run wdidle3 on that drive two days ago stopping the nonsense.

but my original mail had a different target really. I have to recognize that I don't know much about the inner-workings of a filesystem, but I find it odd that once there is no input from the outside, a process keeps writing to the drive indefinitely. in my very narrow thinking the fact that these writes dissapear after a remount would prove their redundance.

to wrap it up, I see no logic to the above and this is why I ask the list to tell me if this is 

 a. something that can easily be fixed via an option I failed to find
 b. a critical part of xfs's internals that cannot be 'disabled' (with a short explanation)
 c. simply a bug

with the little side-story with the WD 20EARS i was just portraying where this default behaviour can get to. 
I don't usually read marketing material, but WD acknowledges that their green drives are wrecked in Linux and they simply encourage customers to change their OS. I just have to ask why have we got to get to this point.

the drive I was trying to get into standby in the first half of my mail is a different one, an enterprise ST31000340NS placed on an otherwise low power ProLiant MicroServer. Having 8W*12h*30 = 2880 Wh less to pay per hdd per month would be easily achievable if the standby mode would be reached when possible.

cheers,
peter