[mythtvnz] HDD Load Cycle Count ticking over too fast

[David Moore]

On 26/06/12 11:51, David Moore wrote:
> My drive is at least a year old and has only reached an LCC of 24,000 
> but is ticking over every 15s now. I guess it hasn't been idle that 
> much. Either playing/recording or turned off. OR maybe it's something 
> new since I upgraded from myth 0.23 to 0.25 which would explain how a 
> year old disk is only up to 24,000 LCC. I wonder if 0.25 is 
> periodically polling/probing all storage disks?
>
Nope. It's not myth diddling with the drive. Killed the backend and the 
LCC still ticks over.

This stops it:

while true; do mv /media/sdc1/file1.zip ~/; sleep 5; mv ~/file1.zip 
/media/sdc1/; done

Home dir is not on the problem drive which is sdc1. file1.zip is about 36M.

Also some googling reveals two Linux utils to set the WD idle timer:

http://idle3-tools.sourceforge.net/
http://sourceforge.net/projects/hdparm/

 From this thread: http://www.gossamer-threads.com/lists/mythtv/users/500147
it seems hdparm uses the idle3-tools code. The thread also talks about 
an interesting backup tool.

But I'm not convinced changing the idle timer to a large value is the 
best solution since the drive will still eventually spin down and then 
immediately spin up again within 15s. I suppose I could stop intellipark 
altogether. I'd really like to figure out why the disk gets pinged every 
15s when there is nothing that should be reading from or writing to it 
unless a myth recording or playback is in progress. Or am I missing 
something fundamental about how Linux caches file system info even on 
idle file systems?