Very high Load_Cycle_Count (over 5 000 in less than 180h)

[Myron]

Does make one wonder if WD have configured the drive to park every 8 seconds on purpose to kill the drive after about 3 years but like any good politician deny that this is the case but is meant to save power. I've had this treatment before. Worst case (gods honest truth!) I wrote to the secratary of the CEO of Fujitsu and then a week later the issue I had (dead drive died under warranty) I had a reluctant call from the London office of Fujistu. Thry had been instructed from the top to resolve the problem by sending a new hard drive, but was told I should have called the London office and not gone that high up...  Erm...  I DID CALL THE LONDON OFFICE AT THE START!!  

I'm gonna mose my hair if I'm not careful and to stress busting excercises!

[richUK]

The problem with all this is that if the drive does die in the next few months you will never know if its down to WD or the fact that you have played around with it. 

Later.

http://community.wdc.com/t5/My-Book-Live/LED-problem/m-p/185624/message-uid/185624/highlight/true#U1...

[Myron]

 jazzymood, I'm chuffed to bits.  Thanks. LED's still working, erred on side of caution and also set the time to intellipark to 8.5 minutes.  smartctl showing a nice stable Load_Cycle_Count.  It's stopped the count to dooms day!  :-)   (Remembered to shut down the MBL, pull the power and re-insert.)

Within 250 hours of operating time the Load_Cycle_Count got to 8326.  That's just down right crazy.  What would be the right value to use for `-s` to set the intellipark timeout to 4 minutes?

I'm assuming this drive, like many a few others I know will panic park the heads if the drive experiences a shock?

[TonyPh12345]

It's a Western Digital Green Power drive in there.

I don't think they have G-Shock sensing in those.

[Myron]

Gotya. Thanks. The idle3 tweak's worked. very well indeed and theunit still goes to sleep after 10 minutes, which I configured it too. The NAS is physically located in a corner of the room where no-one really goes so I'll just have to give it a dust down now and again.  The G-shock sensing was a question born from curiosity.

[jazzymood]

For finding out custom timer values use this tip:

Enter some random value with

./idle3ctl -s 115 /dev/sda

 and then use 

./idle3ctl -g103 /dev/sda 

 It will show your set value in seconds.

That way by changing -s value and checking with -g103 you can find what is right value for what you want.

Edit:

I CAN CONFIRM THAT LOAD COUNT FIX WORKS ALSO ON WhiteLight edition of MyBook (you need to compile it directly on MyBook World Edition using 'make' ).

I set same value as for Live Edition and its working. Awesome!!!

[Myron]

I've made a educated compromise.

NAS:~# idle3ctl -s 141  /dev/sda
Idle3 timer set to 141 (0x8d)
NAS:~# idle3ctl -g103 /dev/sda
Idle3 timer set to 390.0s (0x8d)
NAS:~#

 I'm guesstimating that if the drive does not receive any requests for more than 6.5 minutes then there is a high probability that it's not going to be used for a while so let the intellipark feature park the heads and the load/unload of the heads wil be negligable. The default of 8 seconds programed in by WD was just mad and going against common sense.

Makes sense. If the heads are parked then there is less chance of damage to the heads and or disc.  The NAS's power management can then decide when to spin-down and spin-up the disc. It's also quicker to just un-park the heads as opposed to spinning-up the disc which would cause additional stress to the mechanical and electrical components of the drive.

I have noticed that with the idle3 value set to 145, the drive's operating temperature is higher by about 7 degrees centigrade.

The model of the drive in my MBL NAS is:

Device Model:     WDC WD10EARS-003BB1
Firmware Version: 80.00A80

Dependig on how `193 Load_Cycle_Count` I might try allowing intellipark to ark the heads sooner nut NOT after 8 seconds.  Maybe about three minutes?  Possibly.  6.5 minutes seems fine at the moment.  The load cycle count is not incrmenting at the same rate as the national debt!  Just staying at the one value for well over two minutes.

[jazzymood]

Nice research about temperature raising (it was expected to rise of course).

Becouse of that maybe I will change value to 141 too

[Myron]

Still researching...  Seems, at times, processes in the OS still touch the contents of the drive so the idle counter gets reset before reaching 6.5 minutes so the heads ain't parked until the 10 minute sleep is reached and the drive is spun-down.  Mind you...  I've located my MBL NAS in another room that's always cool and does not get direct sunlight during the day.

Might just leave it alone now. I'm guessing the temperature of 47 degrees centigrade is well within design limits. My last test with the idle3 value set to 141 has not gone above 41 degrees centigrade. Stands to reason in keeping a NAS that uses convection for cooling in a cool environment.

Mind you...   Is the idle3 timer set to 6.5 minutes or 14.1 seconds?  This bit confuses me a bit.

As to head parking, not sure about the WD10EARS-003BB1 but I've seen a few drives where when the heads park, the heads are mechanically moved away from the disc within the head parking zone so when the disc stops spinning that the heads do not come in contact with the disc as it's only a layer of air created by the spinning disc that keeps the heads floating above the disc platter.

If this is the case then constant parking is, physically and mechanically, bad for the drive. Comes back to the conspiracy theory that many things produced now have a MTBF deliberately built into the design and I certainly want a better fighting chance of the WD10EARS-003BB1 not packing up after about three years taking all my files with it. 

[Myron]

Some interesting chatter on this issue on a more generic forum all about storage.

http://forums.storagereview.com/index.php/topic/27269-wd-greenpower-drives-head-parking-issue/