You are not logged in.
Hi guys,
Here is how a well known "spin down" problem is presented in our wiki.http://wiki.archlinux.org/index.php/Lap … wn_problem
Suggested solution is easy turn off APM with hdparm -B 255 device or make it least active with hdparm -B 254 device. Drawbacks of this solution:
1. Consumes lots of power.
2. HDD becomes a heated iron at high io.
After playing around with parameters I came to
hdparm -B 128 -S 60 device
With these parameters I can't HEAR any clicking and the temperature of the disk is not so high. But as usuall it doesn't mean anything, and now I just started to gather statistics for this regime.
Statistics for a previous regime (-B 254) was
Power_On_Hours: 5136 - 3393 = 1743
Load_Cycle_Count : 145093 - 136541 = 8552
Load_Cycle_Count / Power_On_Hours: 4.906
I think one week would be enough for an estimation of the increasing rate of Load_Cycle_Count. Is anyone else interested in a similar testing or may be already did one? I think that the wiki page should be modified with more optimal settings (if they are).
Last edited by ogronom (2009-08-13 13:31:39)
Offline
Setting the advanced power management to 128 _should_ stop the disk from parking but an idle timeout of 5 minutes may be too low unless you are on battery and need to save as much as you can.
R00KIE
Tm90aGluZyB0byBzZWUgaGVyZSwgbW92ZSBhbG9uZy4K
Offline
Ok, maybe. What is the reasonable setting of timeout? (at least what are your settings)
My current rate is 3.25 (which is lower than was before), but again it is too early to talk about reliable information.
I don't care so much about power consumption (my current 3.5 hours is enough for my needs). Temperature was "killing" me.
Offline