TR Forums

Thu Dec 15, 2011 10:26 am

Hmm, my Abit IP35 Pro is set to AHCI and wdidle did exactly what it was supposed to on all three WD drives attached to the mobo.

PeregrineFalcon · Thu Dec 15, 2011 10:50 am

just brew it! wrote:
If wdidle doesn't seem to work on all of the Greens, perhaps someone will come up with a background service to touch the drives every couple of seconds to keep the heads from parking. Should be a pretty trivial thing to write.

Shouldn't take more than a batch file. Throw it in a folder you've got write access to so UAC doesn't block it, and run it as a scheduled task on boot.

@echo off

:BEGIN

if exist wait.txt del wait.txt
ping 127.0.0.1 > wait.txt
dir wait.* > wait.txt

GOTO BEGIN

The create/delete should force the cache to commit them to disk. But really, try to get wdidle working.

(Or stop buying WD Greens.) :wink:

LoneWolf15 · Sat Dec 17, 2011 11:52 pm

just brew it! wrote:
If wdidle doesn't seem to work on all of the Greens, perhaps someone will come up with a background service to touch the drives every couple of seconds to keep the heads from parking. Should be a pretty trivial thing to write.

I seem to recall there was a way to jumper the drives to prevent this. I'll do a little research.

EDIT: No, it is usually the WDIDLE command as listed here --Windows Home Server users disable it so it doesn't cause problems on the EARS Caviar Green drives. Apparently newer Greens don't do things the same way and may not need this.

http://www.mediasmartserver.net/forums/ ... 9f4#p80440

FuturePastNow · Thu Dec 22, 2011 5:47 pm

FuturePastNow wrote:
I was dimly aware of the head-parking issue but never paid much attention, despite the fact that I'm using six Caviar Green drives in my Windows Home Server. You've prompted me to remote in and see what their Load Cycle Counts are. I assume the Raw values are hexadecimal:

HD0: EADS: 200: 00000000005B = 91

HD1: EADS: 73: 00000005D74A = 382794

HD2: EACS: 200: 000000000074 = 116

HD3: EADS: 200: 000000000060 = 96

HD4: EARS: 183: 00000000CAD4 = 51924

HD5: EARS: 171: 00000001558A = 87434

It's been 18 days, I decided to check the drives again.

HD0: 00000000005B = 91

HD1: 00000005FE79 = 392825

HD2: 00000000000D = 13

HD3: 000000000060 = 96

HD4: 00000000EB24 = 60196

HD5: 000000017BD1 = 97233

So two of the drives' Load Cycle Counts went up nearly 10,000, two didn't change at all, and one actually went down. I don't know what to make of that.

Thu Dec 22, 2011 5:55 pm

I don't know what to make of that either.

FuturePastNow · Thu Dec 22, 2011 6:58 pm

Never mind. I must have copied the wrong line from the SMART info, or else it fixed itself (but far more likely the former). It's 74 = 116, or unchanged from three weeks ago.

Fri Dec 23, 2011 9:24 am

FuturePastNow wrote:
...

So two of the drives' Load Cycle Counts went up nearly 10,000, two didn't change at all,

Looking at the numbers again, *three* of the counts went up a lot.

Looks like you'd better apply the firmware patch or run something to "touch" the drives every few seconds to prevent the load/unload cycle from occurring.

If WD was smart, they'd modify the firmware to adapt to the usage environment. It really shouldn't be that difficult for the drive to notice that it is being asked to reload the heads right away nearly every time it unloads them, and automatically increase its unload timeout value.

Wester547 · Tue Jan 10, 2012 12:11 am

@just brew it, do you think that by totally disabling APM, that a mechanical HDD could still be worse for wear quicker, given that APM may give allowance for a drive to spin down and park its heads instead of constantly spinning whenever on?

@mako, if you don't mind me asking, what commands exactly did you use to try to disable the timer? I think you can either use /D as a way of "disabling" it or /S [number] to make it a select amount of time before the heads park, but I'm wondering if something like 1,000 would work or if there's a cap like 62 minutes (which I believe /D activates, but I say this wondering if it's possible to reach avoidance of head parking completely through wdidle, save by restarting and/or powering up and down).

Tue Jan 10, 2012 12:31 am

Wester547 wrote:
@just brew it, do you think that by totally disabling APM, that a mechanical HDD could still be worse for wear quicker, given that APM may give allowance for a drive to spin down and park its heads instead of constantly spinning whenever on?

Potentially, yes. But if you happen to have a system which aggravates the head parking issue, and you leave the timer at its default (short) setting, you're going to exceed the manufacturer's spec for head load/unload cycles *long* before you are in danger of wearing the rest of the drive out.

TR Forums

WD Caviar Green head parking

Re: WD Caviar Green head parking

Re: WD Caviar Green head parking

Re: WD Caviar Green head parking

Re: WD Caviar Green head parking

Re: WD Caviar Green head parking

Re: WD Caviar Green head parking

Re: WD Caviar Green head parking

Re: WD Caviar Green head parking

Re: WD Caviar Green head parking

Who is online