Is there a problem with head parks on WD Green HDDs?

AlpineCarver · Post by **AlpineCarver** » Wed Feb 25, 2009 9:50 pm

can somebody explain why wdidle3.exe isn't freely downloadable from WD's website?
i'm not willing to run a .EXE whose origin i know nothing about except an IP address!

i've been a WD customer for years, but their stonewalling on this issue is souring my perception of them as a vendor for future purchases, especially since they acknowledge the problem and provide solutions for buyers of the RE2 version of the drive, but completely stonewall the rest of us, to the point of actually withrawing a tool that can help fix the problem!

zds · Post by **zds** » Thu Feb 26, 2009 5:11 am

AlpineCarver wrote:completely stonewall the rest of us, to the point of actually withrawing a tool that can help fix the problem!

If I had to guess, I'd say the reason is that the EACS firmware is bugged beyond what wdidle3 is able to fix.

That, naturally, does not justify not releasing fixed firmware for EACS drives but makes it even worse.

jznomoney · Post by **jznomoney** » Thu Feb 26, 2009 5:42 am

I can't find the utility wdidle3.exe. Does anyone know a current site to get this utility from? The ftp posted does not seem to work.

ArTiKO · Post by **ArTiKO** » Tue Mar 03, 2009 12:10 pm

ok ok... new problem on wdc green power...

my wdc my book home edition with wdc WD10EACS when is in idle the header continue seek....

I disconnect any data cable (usb or firewire or esata....) only power cable.... the hard disk after power on the disk (after true rotation disk) come to move header in seek.

What is the problem?? Intellyseek tecnology never wok with this problem... My disk is intelligent and found it self the solution of the problem???

great solutions... or not???

Why this disk have this seek continuos?

trxman · Post by **trxman** » Tue Mar 10, 2009 6:31 am

my question to the WD support:

Hello.

I've been using WD drives for long time. Right now, at my home only, I have 3 WD6400AAKS in my desktop and 8 WD7500AAKS in my server.

I'm looking for upgrade for my server. WD20EADS looks nice and I've planed to buy 8 to 10 of them.

And then, I've found this topic:
viewtopic.php?t=51401

As I run my server 24/7 and it usualy has uptime of a year or so, Load_Cycle_Count problem can be very destructive for me. Load_Cycle_Count could very quickly increase in my Linux environment set to buffer and write data in 30 seconds.

So, does this Load_Cycle_Count issue applies to WD20EADS, too?

Thank you in advance.

WD's support answer:

Thank you for contacting Western Digital Customer Service and Support.

We apologize for the inconvenience, however we are not responsible nor can we confirm or deny ANY third party comments made on web page forums.

Western Digital technical support only provides jumper configuration and physical installation support for hard drives used in systems running the Linux/Unix operating systems. For setup questions beyond physical installation of your Western Digital hard drive, please contact the vendor of your Linux/Unix operating system.

Here is the link for the specifications regarding the WD Caviar Green / GP internal hard drives: http://support.wdc.com/product/kb.asp?g ... EADSâŒ©=en

Thank you.

my update to question:

OK, forget about forum, pure technical questions are:

Does WD20EADS unloads heads after 8 idle seconds like WD10EACS does?

Can this power feature be turned off via SMART parameter?

WD's anwer:

Thank you for your reply.

Please see the drive specifications for these two drives,
long-ass URL
Besides the capacity of the drives, the difference between the WD20EADS and the WD10EACS is the following:

The WD10EACS has a buffer speed or CACHE of 16 MB and the WD20EADS has a buffer speed (CACHE) of 32 MB.

Other than that they are similar drives with different capacities with the 8 second same idle time.

"Can this power feature be turned off via SMART parameter?"

Yes, HOWEVER you have to know how to do it and unfortunately we cannot support the drive being used in this manner.

Western Digital technical support only provides jumper configuration and physical installation support for hard drives used in systems running the Linux/Unix operating systems. For setup questions beyond physical installation of your Western Digital hard drive, please contact the vendor of your Linux/Unix operating system.

at last, they've admited the same 8 seconds unload cycle.

now, the interesting thig is:

Yes, HOWEVER you have to know how to do it and unfortunately we cannot support the drive being used in this manner.

anyone succeeded in this?

rseiler · Post by **rseiler** » Sat Mar 14, 2009 1:06 pm

jznomoney wrote:I can't find the utility wdidle3.exe. Does anyone know a current site to get this utility from? The ftp posted does not seem to work.

Working link:
http://home.arcor.de/ghostadmin/wdidle3_1_00.zip

trxman wrote: at last, they've admited the same 8 seconds unload cycle.

now, the interesting thig is:

Yes, HOWEVER you have to know how to do it and unfortunately we cannot support the drive being used in this manner.
anyone succeeded in this?

Yes, it was mentioned up the thread and elsewhere. Here's the usage:

WDIDLE3 Version 1.00 for DOS

DESCRIPTION
- DOS Level utility to setup or report the idle3 value.

FEATURES
- Scan for all drives. Non-WD Drives shall only show the model and serial numbers.
- Uses a Vendor Specific Command to set or get the idle3 timer.
- Timer can be set from 100 ms to 25.5 seconds, in 100ms increments.

USAGE
WDIDLE3 [/S[<Timer>]] [/D] [/R] [/?]
where:
/S[<Timer>] Set timer, units in 100 milliseconds (1 to 255). Default=80.
/D Disable timer.
/R Report current timer.
/? This help info.

trxman · Post by **trxman** » Sat Mar 14, 2009 1:15 pm

thx a lot!

hope this will work on WD20EADS, too!

trxman · Post by **trxman** » Sun Mar 15, 2009 2:33 pm

Code: Select all

C:\>wdidle3.exe /R
CauseWay DOS Extender v3.49 Copyright 1992-99 Michael Devore.
All rights reserved.

Exception: 0D, Error code: 0000

EAX=02AA0060 EBX=02AC0060 ECX=00000001 EDX=00002AA0 ESI=02AA0060
EDI=02AC0094 EBP=02A7F994 ESP=02A7F964 EIP=02A0F604 EFL=00013206

CS=027F-FD600000 DS=0287-FD600000 ES=0287-FD600000
FS=0000-xxxxxxxx GS=028F-xxxxxxxx SS=0287-FD600000

CR0=00000000 CR2=00000000 CR3=00000000 TR=0000

Info flags=00008018

Writing CW.ERR file....

CauseWay error 09 : Unrecoverable exception. Program terminated.

C:\>

no go on win XP SP3

zds · Post by **zds** » Sun Mar 15, 2009 3:33 pm

trxman wrote: no go on win XP SP3

It's a low-level DOS program, like BIOS updaters and all those.. you need to run it from *DOS*, not from Windows. Time to dig out your DOS boot media.

If you lack any, download FreeDOS, it's free in both senses and works like a charm.

trxman · Post by **trxman** » Mon Mar 16, 2009 3:41 am

booting of FreeDOS iso image gives:

Code: Select all

isolinux: Disk error 80, AX = 4265, drive 9F

Boot failed: press a key to retry...

I'll have to take off the hdd and put it in some other computer for purposes of setting idle time.

rseiler · Post by **rseiler** » Mon Mar 16, 2009 11:25 am

If it needs "real" DOS, then there are plenty here:
http://www.bootdisk.com/bootdisk.htm

Rebellious · Post by **Rebellious** » Wed Mar 18, 2009 5:58 am

I had the same problem with a couple of 2.5" WD Scorpio Black WD3200BEKT and fixed it with the bootleg WD utility. The parameter change is not volatile, and it works regardless of OS:

viewtopic.php?t=52626

the name of the utility is ironic, wdidle3

lunadesign · Post by **lunadesign** » Wed Mar 18, 2009 1:15 pm

With regards to the WD10EACS vs WD10EADS question, it appears that the latter does not have the same issue.

I have 4 EACS drives and 2 EADS drives, all used in removable drive carriers as backup drives in the same system. Each one is only powered on for 1-2 hours per week except one which is on a bit more often.

WD10EACS-00ZJB0 #1: 195 load cycles in 23 hours = 8.5/hour
WD10EACS-00ZJB0 #2: 232 load cycles in 21 hours = 11.1/hour
WD10EACS-00ZJB0 #3: 7208 load cycles in 803 hours = 9.0/hour
WD10EACS-00ZJB0 #4: 1164 load cycles in 58 hours = 20.1/hour

WD10EADS-00L5B1 #1: 22 load cycles in 49 hours = 0.45/hour
WD10EADS-00L5B1 #2: 56 load cycles in 87 hours = 0.64/hour

What's interesting is that I generally don't run SpeedFan so I'm not sure why I'm seeing so many load cycles per hour. Should I be concerned?

zds · Post by **zds** » Wed Mar 18, 2009 7:28 pm

lunadesign wrote:With regards to the WD10EACS vs WD10EADS question, it appears that the latter does not have the same issue.

Thank you very much, this was very useful piece of information!

Trav1s · Post by **Trav1s** » Wed Mar 18, 2009 7:37 pm

zds wrote:
lunadesign wrote:With regards to the WD10EACS vs WD10EADS question, it appears that the latter does not have the same issue.
Thank you very much, this was very useful piece of information!

Yes it is

lunadesign · Post by **lunadesign** » Wed Mar 18, 2009 7:42 pm

Trav1s wrote:
zds wrote:
lunadesign wrote:With regards to the WD10EACS vs WD10EADS question, it appears that the latter does not have the same issue.
Thank you very much, this was very useful piece of information!
Yes it is

You're both very welcome.....I'm hoping the WD20EADS follows the trend and doesn't have the issue either. Can anyone with a WD20EADS confirm or deny this?

rseiler · Post by **rseiler** » Wed Mar 18, 2009 8:46 pm

zds wrote:
lunadesign wrote:With regards to the WD10EACS vs WD10EADS question, it appears that the latter does not have the same issue.
Thank you very much, this was very useful piece of information!

Though the EADS still defaults to every 8 seconds. Couldn't that in itself be a problem in the right context?

That amount of time might be appropriate in a laptop (not that this drive is for a laptop, obviously), and appropriate for a backup drive scenario where the disk isn't touched much, but what about as a main drive?

I guess I'm unclear on when it should be left alone, upped to 25.5, or simply disabled--and also why EADS is less of a problem than EACS.

lunadesign · Post by **lunadesign** » Wed Mar 18, 2009 9:18 pm

rseiler wrote:Though the EADS still defaults to every 8 seconds.

Do we know that for a fact? If so, I'm curious why my EADS drives aren't showing the higher load cycle counts that my EACS drives are.

rseiler · Post by **rseiler** » Wed Mar 18, 2009 10:02 pm

lunadesign wrote:
rseiler wrote:Though the EADS still defaults to every 8 seconds.
Do we know that for a fact? If so, I'm curious why my EADS drives aren't showing the higher load cycle counts that my EACS drives are.

I'm too exhausted to re-read the thread, but I'm pretty sure it was mentioned back there. Plus, a quick Google came up with this post (search the page for "25.5"), which looks convincing:
http://www.networkedmediatank.com/print ... ?tid=16327

I just don't see how this is desirable behavior outside of a laptop. Or how it wouldn't take a severe toll on the drive and our nerves.

Tested my WD10EADS last night.
after approx 8 seconds you hear the screech sound(park)
then when you do something you hear it again(unpark)

glassdarkly · Post by **glassdarkly** » Thu Mar 19, 2009 1:54 am

whiic wrote: The attribute we are discussing is
(C1) Load Cycle Count 200 200 0 64 Ok
and in your drive it appears to be the same as Start/Stop Count. This means that either unload feature is disabled in WD10EADS-00L5B1 or (like with WD10EACS-00D6B0) it's simply LYING.

The only way to determine if your drive is affected is to listen for load/unload clicks (they are distinguishable from regular seeks). SMART data is reliable method of determining load/unload count only for WDxxxxCS-00ZJB0.

This was mentioned earlier in the thread. I believe the consensus was that WD did something to change the way the load/unload cycles were recorded in the SMART data for the EADS drives after multiple complaints about the EACS drives.

Can anyone confirm if wdidle3.exe completely eliminates the issue? Someone was saying earlier in the thread that the load count was still increasing faster than it should have been after using wdidle3.

__Miguel_ · Post by **__Miguel_** » Thu Mar 19, 2009 2:02 am

rseiler wrote:I guess I'm unclear on when it should be left alone, upped to 25.5, or simply disabled--and also why EADS is less of a problem than EACS.

I just hope WD isn't cheating on the whole "park-unpark" count with the EADS series...

However, one thing just came to me... Why would the park/unpark head movement increase start/stop cycle count in the first place? Bear with me a little here, please:

1) The harshest thing an HDD can perform is a full spin-up procedure from power down or sleep modes, right?

2) Park/unpark movements, though, only pertain to the data heads, which are expected to be moving CONSTANTLY; or, if not moving, the actuator must hold them in place.

3) This means that head parks only mean the head moved to another place, one that is actually safer (no head crash risk) and more energy efficient (less drag on the platters).

4) Yes, head parks are a little more dangerous in the sense that the arm can break while loading or unloading (not likely, I assume), and yes, data access will have a slight lag if the head is parked.

However, in the end, I don't think there is a significant increase in drive wear because of constant parking, which would explain why

1) We aren't seing EACS drives failing in droves - these probably had a slight firmware "hickup" that incorrectly counts parks as full drive stops (which is yucky, and WD should fix that, since EACS owners no longer get to understand if there is a problem with drive wear derived from start/stop cycles)

2) EADS drives still exibit the same low timing for head park, but the SST counter doesn't increase with them - if my thoughts are correct, then WD wouldn't be cheating at all.

So, besides the whole "we can't predict the health of our drive from the SST counter" issue, do you think what I said makes sense, and WD drives are actually safe to use? I mean, these aren't built to be system drives, they are meant to be used as low-power storage drives, with few occasional acesses. Which means the 8-second park timer isn't actually that bad, since it saves a lot of energy both from the arm actuator and as drag (or lack thereof, in this case).

The only thing not really explained in this angle is why SpeedFan causes heads to unpark like crazy...I mean, it's not like SpeedFan tries to write to each and every drive in the system at each SMART reading, right?

Any thoughts?

Cheers.

Miguel

zds · Post by **zds** » Thu Mar 19, 2009 4:36 am

rseiler wrote: Though the EADS still defaults to every 8 seconds. Couldn't that in itself be a problem in the right context?

I suspect the EACS firmware was just plain faulty, so it was not just timer setting but a real bug (or the combination of these two) that caused high load/unload counts.

What makes me believe this is that 1) count increase even when the load/unload is disabled with wdidle3.exe and 2) they increase even when setting OS side timings to very minimum (and lot lower than 8s).

trxman · Post by **trxman** » Thu Mar 19, 2009 5:13 am

zds wrote:What makes me believe this is that 1) count increase even when the load/unload is disabled with wdidle3.exe and 2) they increase even when setting OS side timings to very minimum (and lot lower than 8s).

and that sounds like a extremely big problem. how come thay just don't want to change that firmware...

lunadesign · Post by **lunadesign** » Sat Mar 21, 2009 12:01 am

Update on the WD10EACS vs WD10EADS comparison

Following up on rseiler's post, I did some additional testing and confirmed that WD10EADS does have the same parking behavior as WD10EACS except that just like the WD10EACS-00DB0, WD10EADS isn't updating the Load Cycle Count counter. (At least the one I tested isn't.)

I tested my WD10EACS-00ZJB0 first and with SpeedFan running noticed that every minute I'd hear a certain noise, followed by another one exactly 8 seconds later, with the cycle repeating every minute. In this case, the Load Cycle Count variable was incrementing by one during each of these minute long cycles.

Once I knew what to roughly listen for, I tested against my WD10EADS-00L5B1 and heard the exact same pattern (although the two noises sounded quite different) but didn't see the counter increment.

I also had my power meter hooked up and noticed that during the 8 second period, the system was drawing 1.7-1.8W more.

So I guess its safe to say that this is a consistent behavior with all GreenPower drives. And I can definitely see how this parking behavior is critical to the drive's lower power consumption numbers.

The only question that remains for me is why are my drives getting so many load cycles if I'm not running SpeedFan normally?

rseiler · Post by **rseiler** » Sat Mar 21, 2009 6:57 am

luna, thanks for that update.

That's a good use for Kill-A-Watt: I wouldn't have thought parking the heads accounted for that much.

Looking at your earlier post (very high load cycles count for EACS and relatively low for EADS), what environment are you running under? The article below, which is not about either model, mentions Linux in addition to SpeedFan, which you've excluded or only use lightly.

http://www.silentpcreview.com/Terabyte_Drive_Fix

If it's Windows, you might be able to tell something by running Process Monitor and focusing in on events which seem to coincide with audible changes in the drive.

While it may be comforting that the SMART load/unload counter doesn't balloon with EADS, as someone suggested earlier, we don't know if WD implemented a fix whereby it's merely masking what's still happening instead of actually changing its behavior (which on the face of it looks exactly the same as with EACS).

lunadesign · Post by **lunadesign** » Sat Mar 21, 2009 9:09 am

rseiler - the system with these drives is running Windows XP SP3.

Most of the drives are backup drives used with EMC Retrospect. They only get powered on once a week for an hour or two to receive data and then are powered off. When they're on, they're generally used non-stop although its possible that there are > 8 second pauses while the backup software is doing some thinking. I'll watch this carefully next time the backups run on an EACS drive.

The other ones are archive drives that contain seldom used very large files (Ghost images, virtual machines, etc). These only get powered on about once a week for anywhere from 2 hours to 8 hours. The files they contain are generally copied to/from the drive over the local network using either Windows Explorer or Beyond Compare (a great file/folder comparison tool). This usage pattern is generally a few large file copies or comparisons with a few hours of idle time in between. I'm going to try to do some testing with an EACS drive to figure out what's triggering the load cycles.

I'm convinced that the EADS is just like the EACS but masking the numbers, just like some of the EACS units apparently do. The bummer is that this masking makes it much harder to figure out what system usage is causing the extraneous load cycles since we can't use SMART to monitor it. Which is really strange because WD's recommendation in the KB article (http://wdc.custhelp.com/cgi-bin/wdc.cfg ... 1232125802) for this issue for RE GP drives is to adjust your OS or applications to stop waking up the drive too often. Its kinda hard to do that if they mask the SMART counter that tells us this is happening. Sheesh.

zds · Post by **zds** » Sat Mar 21, 2009 5:07 pm

lunadesign wrote:And I can definitely see how this parking behavior is critical to the drive's lower power consumption numbers.

The bad thing is.. apart from the power comsunption, this kind of behaviour is not what people at this site (SPCR) want. Constant loading/unloading, while it might lower the absolute power usage does not help in making the drive silent.

karazy · Post by **karazy** » Sun Mar 22, 2009 5:31 pm

I have a WD15EADS 1.5TB Green Drive and wdidle3 reports that the idler is set for 8 seconds. Smart monitoring shows LCC on par with PCC even though intellipark does exist so it is possible that the newer Western Digital drives now mask LCC caused by intellipark which makes it harder for us to tweak our OS's or even do any kind of testing.

I'm wondering what effects will disabling intellipark have. I know it will use up more energy and mostly likely generate more heat but to what extent? How much more power will it use up, how much more heat will it generate? Will disabling intellipark simply turn my hard drive into a non-Green drive with little or no negative impact in performance and integrity?

Benjiro · Post by **Benjiro** » Sun Mar 22, 2009 6:09 pm

Try running those WD Green's in a 4 drive, Raid 5 set. Then you will notice that head parking even more as the 4 hd's need to unpark there heads.

karazy:

Turning off intellipark in theory will increase the response time as they don't need to unpark every time you access your system ( with no activity ). But, it will not increase the general response time for any actions that happen in that 8sec interval... So don't expect that 5400rpm to suddenly turn into a 7200rpm drive. Unlike the misleading advertisement that WD released, all those green drives are 5400 rpm, and there is no "intelligent" speed control system

I'm surprised nobody sued them for false advertisment with there "up to 7200rpm" ****.

To be honest, whats the use off intelpark? You can just as well put your HD's into sleep after 1min off inactivity, and have a bigger power save... I suspect that the constant parking & unparking is also going to draw "extra" power.

Personally i think that data integrity is not going to be harmed. When a drive dies, most off the time its during the system startup.

The drive heats up, the materials expand, the motor off the drive is going to need more voltage, etc... On shutdown you have the reverse effect, as the drive cools down.

Do not underestimate the effect of material fluctuation because off heat-up, cool-down, and power spikes from the startup... One off the reasons i'm a firm believer in keeping the system running 24h.

Personally i'm going for a few nice 7200rpm 1.5TB HD's next. The power difference is not worth the general discomfort, and lower speed/response from the lower rpm, head parking...

Per Hansson · Post by **Per Hansson** » Tue Mar 24, 2009 10:16 am

Just got a somewhat eery message from WD; The wdidle3 util will void the warranty if used on RE2-GP drives! (thus I have not used it)

Here are the messages;
Customer (Per Hansson) 03/16/2009 01:55 PM
Hi, I have updated the firmware on my 3x RE2-GP 1TB drives that I am running in a RAID5 array, I've done this because I noticed that the S.M.A.R.T data showed a very high number of unload cycles of 88462 times (and two of my drives have already failed and are replaced through RMA)

I saw in your PDF document for the update the text that "This timer can be changed to values between eight seconds and five minutes with a utility that will be provided upon request."
I'm just asking if I can get this utility and what you recommend I set the timer to, I have attached 3 pictures of the SMART data from my harddrives

-----------

Dear Per,

Thank you for contacting Western Digital Customer Service and Support.

Please refer to the link below for all the info you requested, it also contains a link to download the tool.

Answer Title: The S.M.A.R.T Attribute 193 Load/Unload counter continue to increase for the WD RE2-GP SATA II hard drives
Answer Link: http://wdc.custhelp.com/cgi-bin/wdc.cfg ... 1232125802

If you have any further questions, please do not hesitate to contact us.

Sincerely,

---------------

Customer (Per Hansson) 03/17/2009 12:15 PM
Hi, as I wrote in my post below I already HAVE updated the firmware, so I
already have used the files you link, what I want is what is referenced in
the PDF on that page you linked me;
"This timer can be changed to values between eight seconds and five
minutes with a utility that will be provided upon request."

Sincerely - Per Hansson, editor @ Techspot.com

-------------------

Dear Per,

Thank you for your response.

Unfortunately The WDIdle Utility is no longer available from Western Digital.
The problem here is that WDidle is a tool which was developed for older drives but still works on some new drives.

It was not designed for this and so the outcome can be unpredictable, hence we do not support it.

It can be found throughout the Internet but is no longer distributed by WD

By using WDIdle your drive's firmware is altered and the warranty will be voided.

Sincerely,

Is there a problem with head parks on WD Green HDDs?

ditto