Laptop drive die prematurely: fix for sabayon/gentoo?

You guys all know about http://ubuntuforums.org/showthread.php?t=795327
but unfortunetly thats for a debian based distro. sabayon is based on gentoo so that fix isnt exactly going to work.

i tried to copy and past the script command line into etc/conf.d/hdpram but either im doing it wrong or that is not how to do it.

anyone here figure out a way on how to fix this problem in sabayon? i tested it to see and it is still happning in sabayon.

 

How are you sure your machine is affected?
Check the HDD, by model doesnt help at all, you need your HDDs model.
In windows, right click the C drive.
It should be the first thing listend under ther hardware tab, with your DVD drive as well.
For example: The Vostro 1000, nearly the exact same notebook as me(well it i, just repainted) is affected(mine isnt listed), but my HDD isn't.

 

I'm not sure what you're trying to paste into /etc/conf.d/hdparm, but that's a configuration file for the hdparm program to get settings from when it loads, not someplace for running a script. If you read this thread over at the Gentoo forums, you see that some there recommend making a change to the /etc/conf.d/hdparm configuration file to alleviate excessive HDD load cycles. (You of course need to run hdparm when you boot by adding it to your default runlevel with the rc-update command as well). So this is an example /etc/conf.d/hdparm file, based on some of those recommendations;

Code:

# /etc/conf.d/hdparm: config file for /etc/init.d/hdparm

# You can either set hdparm arguments for each drive using hdX_args,
# discX_args, cdromX_args and genericX_args, e.g.
#
# hda_args="-d1 -X66"
# disc1_args="-d1"
# cdrom0_args="-d1"

# or you can set options for all PATA drives
pata_all_args="-d1"

# or you can set options for all SATA drives
sata_all_args="-B200 -S252"

# or, you can set hdparm options for all drives
all_args=""

From the hdparm man page;

The thing to be careful about here is to make sure, especially on a laptop HDD, that you do not run the HDD too hot. Make sure you're periodically monitoring the temperature of your HDD both before and after you make these kinds of changes.

Good Luck..

 

This consequences of this issue are disputed, as many people report no hardware failures as a result of spinning down frequently. Other people have pointed out that this doesn't just happen in Ubuntu, but Windows as well. In addition, the fix of keeping the platter spinning all the time isn't that much of a fix, since you're still wearing out the drive. Either way the motor bearings are being overworked.

 

I agree that this "problem" is likely overstated. I've never had problems with HDDs on over a half dozen laptops I've owned, and I've never messed with these settings. I did try the hdparm settings mentioned in my post above, to see what the effect would be, and it subjectively felt like my system was responding slower, and it seemed like my HDD temp had risen slightly during standard operation (1 or 2 degrees), so I changed them back to the default (nothing).

Good Luck..

 

well i did that whole watch load cycles for 15 mins test that is mentioned on the ubuntu page.

 

Another datapoint to back that up. I bought this 7K200 in August 2007:

Code:

$ sudo smartctl -a /dev/sda | egrep -i '(load|power|error)'
					without error or no self-test has ever 
					power-saving mode.
Error logging capability:        (0x01)	Error logging supported.
  1 Raw_Read_Error_Rate     0x000b   100   100   062    Pre-fail  Always       -       0
  7 Seek_Error_Rate         0x000b   100   100   067    Pre-fail  Always       -       0
  9 Power_On_Hours          0x0012   085   085   000    Old_age   Always       -       6610
 12 Power_Cycle_Count       0x0032   100   100   000    Old_age   Always       -       591
191 G-Sense_Error_Rate      0x000a   100   100   000    Old_age   Always       -       0
192 Power-Off_Retract_Count 0x0032   100   100   000    Old_age   Always       -       24
193 Load_Cycle_Count        0x0012   032   032   000    Old_age   Always       -       689413
199 UDMA_CRC_Error_Count    0x000a   200   200   000    Old_age   Always       -       0
223 Load_Retry_Count        0x000a   100   100   000    Old_age   Always       -       0

That's ~100 load/unload cycles/hr on average. As you can see, this drive has never stuttered.

 

Thanks. My 7K200 bought in December of 2007

Code:

smartctl -a /dev/sda | egrep -i '(load|power|error)'
                                        without error or no self-test has ever 
                                        power-saving mode.
Error logging capability:        (0x01) Error logging supported.
  1 Raw_Read_Error_Rate     0x000b   100   100   062    Pre-fail  Always       -       0
  7 Seek_Error_Rate         0x000b   100   100   067    Pre-fail  Always       -       0
  9 Power_On_Hours          0x0012   092   092   000    Old_age   Always       -       3761
 12 Power_Cycle_Count       0x0032   100   100   000    Old_age   Always       -       126
191 G-Sense_Error_Rate      0x000a   100   100   000    Old_age   Always       -       0
192 Power-Off_Retract_Count 0x0032   100   100   000    Old_age   Always       -       30
193 Load_Cycle_Count        0x0012   055   055   000    Old_age   Always       -       452833
199 UDMA_CRC_Error_Count    0x000a   200   200   000    Old_age   Always       -       0
223 Load_Retry_Count        0x000a   100   100   000    Old_age   Always       -       0
SMART Error Log Version: 1
No Errors Logged
If Selective self-test is pending on power-up, resume after 0 minute delay.

Good Luck..

 

You write that your numbers back up the fact that the load cycle issue is overstated.
You've reached more than 600000 load cycles in less then a year and thus exceeded the number of load cycles after which you can expect hard drive failure...
Do you mean that because your HD still runs, the load cycle issue is overstated?

 

Do you see my error counts? This drive is been in two (different) laptops that are hugely mobile, and neither the hard nor soft error counts is anything but zero.

... so, yeah, at least for the Hitachi 7K200.

 

Righty, I'm pretty much a n00b when it comes to HD's. (Although I can read reviews and have a HTS 7K100 myself ).
So it's about the errors read by Smartmontools which should be taken into account instead of focussing on '193' load cycle count?
Could you explain what the different 'errors' are and what the hard /soft error difference is?
I've no idea what's the difference between 1 Raw_Read_Error_Rate, 7 Seek_Error_Rate, 191 G-Sense_Error_Rate and 199 UDMA_CRC_Error_Count.

Thanks.

 

It's also something Bog pointed out earlier up in the thread;

On this Ubuntu Brainstorm thread there's a poster who shares experience to back this up;

Maybe the current best advice is instead of fiddling with hdparm settings, backup the system regularly, buy a backup HDD, and then wait for the inevitable swapout and restore. At least until the HDD vendors and OS vendors sort this one out..

Good Luck..

 

Thanks for your reply Jas!

To be honest, I have tried fiddling with hdparm settings, I managed to decrease the load count cycle significantly without any temperature rise.
I do not know if those settings resulted in wear and tear in another department like you and Bog have pointed out.

Unsatisfied with my uncertainty regarding my HD I finally contacted Hitachi Support and asked them about the load cycle count 'issue'.
I've told them that the agressive power management setting on my Hitachi Travelstar 7K100 combined with Ubuntu/Linux Mint has lead to a high number of load cycles (>120p/h) and asked them if they knew a solution.
First I was told to look at the Ubuntu forums for the hdparm setting solutions there;

Dear Sir,
Thankyou for your email.
This sounds very similar to the recent reports that were seen by people
using the Ubuntu flavour of Linux.
https://bugs.launchpad.net/ubuntu/+source/acpi-support/+bug/59695
In the comments section there are various linux commands that allegedly
solve the issue.
Regards
Adrian Cutler
Hitachi Technical Support Center

I then told them I knew of those 'solutions' but wanted to know if there wasn't a better way of handling the load cycle count issue.

This is the reply I got from Hitachi technical support;

Dear Sir,
Thankyou for your reply.
We have tested this further in our CASLab and would suggest the following :
Download the latest version of the Hitachi Feature Tool - v2.08 - from the
following page :
http://www.hitachigst.com/hdd/support/download.htm#FeatureTool
Create the self booting diskette or CD and reboot your system. Once the
program has started, selct your drive, and from the Features menu, select
'Change Advanced Power Mode'
Move the slider to the right until 'ACTIVE IDLE' appears on the display.
The actual value does not matter, we just need it to show ACTIVE IDLE.
Press the Save button and then shutdown the system. The system needs to be
actually turned off for the changes to take effect, a warm reboot will mean
the changes have been lost.
Now when using in a Linux system the HDD will never unload the heads When
this mode is entered the HDD power requirements will drop by between 45 and
55% so heat generated will also reduce.
The point at which the HDD changes to Active Idle mode is determined by the
Host's access patterns to the disk.
Best regards
Adrian Cutler
Hitachi Technical Support Center

So, according to Hitachi, this is the way to go if you run Ubuntu (or a Ubuntu derivative like Linux Mint).
What do you guys think about this advice?

 

Well they did make the HD. And if it kills your hard drive, im sure you can save the email for proof? Then again you would have to prove that is what did it.

I talked to seagate about it, via live chat. I was told "it was an issue with the hard drive controller, not the drive. We do not support hard drive controllers"

Maybe i will get better results via email.

 

Baserk,

Thanks very much for your post. It cleared up a lot in terms of what Hitachi recommends for drive settings, and the best way to actually set the drive settings. Even before your post I had downloaded the latest version, 2.10, of their Feature Tool from here, and now with your information, I went ahead and booted my laptops with it. What I found for two different laptops, and two different Travelstar 7K200 200gb drives, is that apparently the default power management setting for this drive, (see the hdparm -B option earlier in this thread), is 128, which translates into what Hitachi calls, "Low Power Idle".

So using your email from Hitachi support I moved the slider on the Feature Tool APM settings screen to the right, and discovered that the minimum level setting in the "Active Idle" range is 192. So I changed both drive's APM setting to 192. I found it interesting that while the Linux discussion of this issue has been to try and manage both the drive APM and Spindown settings, the Hitachi email suggested only changing the APM setting. As a matter of fact, there is no Spindown setting, (see the hdparm -S option earlier in this thread), in the Feature Tool that I could discover, so I did not change any values related to that. I also noticed, for whatever reason, that the Acoustic Mode was not set on either drive. I enabled it at the default recommended setting of 128, and then as you described in your email from Hitachi support, I turned off my machines.

Now when I boot either XP or Linux, and check the APM settings of the drives, it's reported as 192. Also the acoustic setting is reported as set at 128 as well. I find the recommendations in the email suitable for my environment, and setting them with Feature Tool, allows my system to reboot into either OS, and I don't have to worry about running hdparm to set the drive every time the system boots, or NHC in XP, for the same thing. So thanks for that help, and I would encourage everyone who uses these Hitachi drives to use the Feature Tool to set the drive settings to whatever setting you want, regardless of not only what OS you run, but also what version of Linux you use. I don't think this is an issue specific to anything other than the APM settings of your laptop HDD.

BTW, I also tried to email Hitachi HDD support and pointedly ask them what settings are recommended, etc. I'm on my third send of the email without receiving a response, so I'm really glad you shared your information from the email they sent you. If they ever respond to my request for information, I will make sure I revisit this thread and post their reply.

Good Luck..

 

Usually, or pretty much always, it's you helping out the linux users here on NBR so I'm glad I've been able to once offer some help to you Jas.
I was just pleased that I could solve this issue on the appropriate hardware level and not have to worry again whenever I change distro's/install XP again/dual boot etc.
Hopefully your explanation can indeed help others too to fix this once and for all.
I'm pretty sure if you wrote a thread on this you'd please a lot of folks, because I don't think many will read about a sabayon/gentoo fix...
Cheers.

 

Thanks for sharing all this info, Jas.

I tried some of the fixes for this "issue" once, and I found that they were merely making the HD platter spin all the time. Not only does this put the wear on some other hard drive component by continuously wearing out the motor bearings, but it also produces significant heat, vibration, and power consumption.