PDA

View Full Version : [ubuntu] Test If Your Laptop Drive Is Going To Die Prematurely - Please Read All Laptop Users



nicedude
May 15th, 2008, 06:17 PM
MODELS OF LAPTOPS THIS IS CONFIRMED TO BE AFFECTING - PLEASE IF AFFECTED POST HERE SO THIS THREAD STAYS BUMPED UP FOR PEOPLE TO SEE - LIST IS NOT COMPLETE SO PLEASE TEST YOUR EQUIPMENT

ACER 5520-5716 - Reported by me
THINKPAD T6 - Reported by reacocard
VAIO VGN-N21E - Reported by playroll
ASUS F3JA - Reported by boppp
Acer Aspire 7520-5839 - Reported by Anduu
HP dv6000 - Reported by ASULutzy
Gateway M-1617 - Reported by johnl
Lenovo thinkpad R61i - Reported by miga
acer model 5920-ZD1 - Reported by beyboo
Hitachi HTS541616J9AT00 - reported by halfmanhalfbug
HP 530 - Reported by Arthur Archnix
Lenove T60p - Reported by Crackh34d
Toshiba Satellite A135-S2246 - Reported by volkswagner
Dell Inspiron 640M/E1405 - Reported by RRFarFar
Toshiba Satellite ProL10 - Reported by webabun
HP nx6125 - Reported by valloric
dell xps1210 - Reported by RRFarFar
Toshiba Satellite A75 S211 - Reported by wdecker
Dell Vostro 1000 - Reported by epitaph
Dell Vostro 1500 - Reported by goldenboy
HP dv6424 - Reported by linuxuser187
HP Pavillion dv9804 - Reported by epitaph
Compaq F572US - Reported by exharrison
Acer Aspire 5570-4315 - Reported by pharnet
HP 6710B - Reported by joshoekstra
Acer Aspire 7520-5A2G12Mi - Reported by Minardi
Gateway M-6846 - Reported by Recruit0
Dell Inspiron 6400 - Reported by Lapino
HP DV6810us - Reported by packman1234
Packardbell EASYNOTE_MX67-O-070 - Reported by Kestol
HP Pavilion dv2058ea - Reported by crisswyatt
Sony Vaio VGN-FZ180E - Reported by G@BO
Asus U6s - Reported by kuyote
Toshiba Satellite A105-S4254 - Reported by darco
Dell Latitude D520 - Reported by llevering
Dell Vostro 1200 - Reported by Super Jamie
HP NX7400 - Reported by Geniusj
Qosmio F40 - Reported by marcon00
Dell Inspiron 1720 - Reported by Cantodea
Zepto Znote 6615WD - Reported by Rokuta
HP nx6325 - Reported by Andycas
Toshiba Portege 4010 - Reported by iamvlad
HP dv 9000 dv9058ea - Reported by qbashi
Dell Inspiron 1520 - Reported by LeoSolaris
Toshiba A135 S2386 - Reported by dirtblack
acer travelmate 4603 WLCi - Reported by ditdot
Znote 3215W - Reported by laoshi
Compaq v5207NR - Reported by strick242
lenovo y510 - Reported by tlcrepairs

DRIVES CONFIRMED BY MEMBERS TO BE AFFECTED
TOSHIBA MK2035GSS HARD DISK - Reported by linuxuser187
Hitachi HTS541612J9SA00 HARD DISK - Reported by epitaph
Western Digital WD3200BEVT-60ZCT0 - Reported by linuxuser187
TOSHIBA MK8037GSX - Reported by exharrison
Hitachi HTS54161 - Reported by pharnet
TOSHIBA MK2546GSX - Reported by joshoekstra
Hitachi HTS541612J9SA00 - Reported by Minardi
WDC WD3200BEVT-22ZCT0 - Reported by Recruit0
TOSHIBA MK2035GS - Reported by G@BO
Seagate ST9160821AS - Reported by kuyote
Toshiba 1234GXS - Reported by llevering
Samsung HM121HC ( 2.5" Hard disk ) - Reported by dmcn
Fujitsu MHY2250BH - Reported by Super Jamie
TOSHIBA MK8034GSX - Reported by Geniusj
Hitachi HTS542512K9SA00 - Reported by marcon00
SAMSUNG HM120JI - Reported by marcon00
Seagate ST9250827AS - Reported by Cantodea
Hitachi HTS721080G9SA00 - Reported by Rokuta
Samsung HM250JI - Reported by Andycas
ST94813A - Reported by Fenris_rising
Hitachi HTS542525K9SA00 - Reported by tlcrepairs
Hitachi-Travelstar 250GB 0A55566 - Reported by strick242
Seagate ST9808210A - Reported by ditdot

THE FOLLOWING LISTS OF AFFECTED SYSTEMS FROM OTHER SOURCES WERE COMPILED AND GIVEN TO ME BY REACOCARD FOR MEMBERS INFO - SO CLICK HIS STAR AS HE DESERVES YOUR THANKS

LIST OF DRIVES THIS IS BEEN REPORTED TO AFFECT
Fujitsu MHW2120BH
Hitachi C4K60 (HTC426060G9AT00)
Hitachi HTS541010G9SA00
Hitachi HTS541060G9AT00
Hitachi HTS541040G9AT00
Hitachi HTS541210H9SA00
Hitachi HTS721010G9SA00
Hitachi HTS541616J9SA00
Hitachi HTS541080G9SA00
Samsung MP0804H 80GB
Samsung HM12HII
Seagate Momentus 7200.1
Seagate ST910021AS
Seagate ST9160821AS
Seagate ST92811A
Toshiba MK1234GSX
Toshiba MK1032GSX
Toshiba MK3006GAL
Toshiba MK2035GSS
Western Digital WD1200VE
Western Digital WD800BEVS

COMPUTERS POTENTIALLY AFFECTED BY THIS BUG
Acer Aspire 1662wlm
Acer Aspire 1642 WLMi
Asus G1S-A1
Dell Inspiron 6400
Dell Latitude C640
Dell Insipron 9400
Dell Latitude C-840
Dell Latitude D830
Dell XPS M1210
Fujitsu P7230
HP compaq nw9440
HP Compaq nx6325
HP Pavilion dv6500
HP Pavilion dv6602
Thinkpad T61
Thinkpad T60
Thinkpad R52
Thinkpad R50e
Thinkpad T43
Thinkpad T42
Thinkpad T23
Thinkpad Z61
Thinkpad T60p
Toshiba P205
Toshiba M55

Thanks Reacocard for the lists & Thanks to Kestol for questioning the activity in first place.

Several links to offsite information about this bug as the one I got from a member and had posted before was dead.

Launchpad
https://bugs.launchpad.net/ubuntu/+source/acpi-support/+bug/59695

Ubuntu brainstorm
http://brainstorm.ubuntu.com/idea/288/

Overclocking Wiki
http://www.overclockingwiki.org/forums/showthread.php?t=1645

So everyone if you are interested in this bug the above links work and can possibly inform you better as to methods other than mine to fix this.


PLEASE POST IN THIS THREAD ANYONE WHO CONFIRMS THEIR LAPTOP IS SUFFERING FROM THIS WITH YOUR MODEL# SO I CAN POST IT HERE FOR OTHERS TO SEE

1. MORE INFO ON THIS PROBLEM IN GENERAL - THIS PROBLEM IS PRESENT IN MANY LINUX DISTROS NOT JUST UBUNTU BUT IT ONLY AFFECTS CERTAIN DRIVES

The problem occurs when APM ( power management ) is constantly parking and unparking your HDD due to aggressive power management and is a hardware firmware flaw present in certain brands & models of drives. If the problem exists for you it will increase your HDD cycle count rapidly and as for my drive it is rated for 600,000 cycles for its lifetime and I would have hit that in 8 months of constant use so please read and tell your friends using any OS to check it out on their laptops as well. I am just trying to save someone having their device bricked by this. I don't get any pay or even glory for my time to warn you so I am doing this out of the goodness of my heart to save you grief down the road. Someone else ( Kestol ) mentioned below asked the question about this abnormality and at first I thought he was wrong ( Sorry Kestol ) then I found that my laptop suffered from it so I made this thread to WARN YOU to test yours as well.

First thanks to Kestol for figuring out this was taking place. I believe through my research of this situation that this warrants enough concern that you should install the simple tool below and test your laptop since if this problem does what it looks like it can, this bug could cause your HDD to prematurely fail. If you don't believe me then I can at least say I warned you and my conscience is clear, plus I won't have to pay for your hardware when it goes poof. But for people who are affected & would rather protect themselves than risk damage, I have made a simple script to fix this problem. First I would recommend to anyone with a laptop to use the software in step 2 to test if your system is actually affected by this problem or not, if not then you don't need this script and you are fine ( Don't use the script if you are not affected ). If so however then I suggest you follow my instructions and implement this fix to stop the disk cycles from growing quickly and possibly killing your hard drive. I can confirm that this is present in Hardy 8.04 so the newest OS does not mean you are not affected, although this is not a Ubuntu flaw but a general OS one ( Google it and see for yourself that it can affect Linux & Windows ) MY DRIVE WAS DOING 120 CYCLES AN HOUR MEANING I WOULD EXCEED THE MANUFACTURERS RATED LIFE EXPECTANCY IN LESS THAN ONE YEAR

-----------------------------------------------------------------------------------------------
This problem is only present due to the existence of the three following factors:
* Hardware is set (default or otherwise) to aggressive power management, causing heads to park. (default behavior of many drives and often the only user available type of power management)
* Disk is touched often, causing heads to unpark. (default behavior of many distributions)
* Drives are spec'd to a limited number of these cycles. (600,000 is the most common, although some may be spec'd higher or lower).

Reasonable Limits / Criteria for a fix:
* There should be fewer than ~15 load cycles per hour, except during heavy usage while on battery.
* This provides a life expectancy of over four years, which is reasonable for a hard disk.
--------------------------------------------------------------------------------------------------------------
The above acceptable statements were taken from other web sources ( Launchpad bug report ) on this problem so I am only doing the best I can to offer advise and support since I hadn't seen anything about this here. If you do the test then multiply the figure you get by 4 ( If you waited 15 minutes ) and that will give you an hourly increase count from there it is up to you if the numbers of hours of operation until your drive will reach its rated lifespan for this activity is acceptable or not as your the one who will have to replace your drive if it dies.

I did some calculations and made these tables so everyone could easily see what their result means to them.

If you have your laptop on 8 hours per day
15 cycles per hour - how long till 600,000 barrier 13.5 yrs
25 cycles per hour - how long till 600,000 barrier 8.1 yrs
50 cycles per hour - how long till 600,000 barrier 3.9 yrs
75 cycles per hour - how long till 600,000 barrier 2.7 yrs
120 cycles per hour - how long till 600,000 barrier 1.6 yrs
200 cycles per hour - how long till 600,000 barrier 1.1 yrs

If you use it 12 hours per day
15 cycles per hour - how long till 600,000 barrier 9.0 yrs
25 cycles per hour - how long till 600,000 barrier 5.4 yrs
50 cycles per hour - how long till 600,000 barrier 2.6 yrs
75 cycles per hour - how long till 600,000 barrier 1.8 yrs
120 cycles per hour - how long till 600,000 barrier 1.1 yrs
200 cycles per hour - how long till 600,000 barrier 8.0 months

If you leave it on almost 24/7
15 cycles per hour - how long till 600,000 barrier 4.5 yrs
25 cycles per hour - how long till 600,000 barrier 2.7 yrs
50 cycles per hour - how long till 600,000 barrier 1.3 yrs
75 cycles per hour - how long till 600,000 barrier 10.9 months
120 cycles per hour - how long till 600,000 barrier 6.8 months
200 cycles per hour - how long till 600,000 barrier 4.0 months

More than 200 cycles per hour and you better fix it as that would be worse than mine was.
Math in these tables isn't perfect as I rounded and multiplied some figure to do it quickly, but its darn close.
--------------------------------------------------------------------------------------------------------------
everyone should be thankful for HalPomeranz & ethanay for the tips on how to test drive heat levels below in this guide. I have tested mine and my drive is showing 42C under high load with my fix in place so as for me all is well and I hope it will be for all of you as well. My only wish is to help this community and I hope this thread furthers that goal :-)
--------------------------------------------------------------------------------------------------------------

2. HERE IS HOW TO TEST IF YOUR DRIVE IS BEING AFFECTED BY THIS SITUATION AND TEST YOUR TEMP AT SAME TIME

INSTALL SMART MONITOR TOOLS - Disk management software, command is following for install.

sudo aptitude install smartmontools

ONCE INSTALLED RUN THIS COMMAND - COPY THE WHOLE LINE AND PASTE IN A TERMINAL AND RUN IT AND IT WILL OUTPUT BOTH A LOADCYCLE COUNT AND TEMPERATURE FOR YOUR DISK
(Thanks to member HalPomeranz for the new command and thanks to member ethany for the original temp check advise and commands that I changed to this )

date; sudo smartctl -a /dev/sda | egrep '(Load_Cycle_Count|Temperature)'

MY OUTPUT AS AN EXAMPLE
193 Load_Cycle_Count 0x0012 083 083 000 Old_age Always - 171239
194 Temperature_Celsius 0x0002 137 137 000 Old_age Always - 40 (Lifetime Min/Max 11/54)

Notice the first line shows my load cycle count ( Look at last number )
And the second shows my temprature for my hard disk. Just write these both down on a piece of paper so we can refer back to them in a minute.

** NOTE
If your laptop disk is addressed as /dev/hda then change the above command to reflect that ( test for this with "sudo fdisk -l" and then look at output )

My sudo fdisk -l output as an example you can see my disk is addressed as sda

Device Boot Start End Blocks Id System
/dev/sda1 * 1 2550 20482843+ 83 Linux
/dev/sda2 2551 3036 3903795 82 Linux swap / Solaris
/dev/sda3 3037 6683 29294527+ 83 Linux
/dev/sda4 6684 19457 102607155 b W95 FAT32

WAIT 15 MINUTES AND RUN THE COMMAND AGAIN

If in 15 minutes the load cycle count number listed in this commands output changes by more than 2-4 then it is probably too much for comfort, Mine with my fix in place has only increased by 8 in 12.5 hours when previous to fix it went up 120 per hour, which for my drive if running 24 hrs a day gave me a life span of about 8 months which is not very acceptable to me :-(

3. IF YOU ARE AFFECTED THEN FOLLOW DIRECTIONS HERE TO USE MY FIX - PLEASE READ FULLY AND CAREFULLY AS INCORRECT IMPLEMENTATION WILL NOT PROTECT YOU

Your hard disk could be called either sda or hda as you should have learned in step 2 so figure out which your drive is called and then download the script that applies to it. Once downloaded to your desktop right click on the script and go to properties -> in the permission tab select to make the file executable ( this must be done with any script you download or make yourself ). Once you have it set as executable and its on your desktop ( I am assuming your desktop is called Desktop as in you have an English install - if not the script will still work but you must copy it to the 4 locations listed in another way) just run each of the four commands below in a terminal window to copy the script to each location it is needed so that it will fix this problem after you reboot.

I have found that a setting of -B 200 works just fine here and my cycle count has gone up 8 in 12.5 hours vs 120 per hour prior to this fix, so I am happy that it is fixed for me. If my script value doesn't work on your machines hard disk ( which so far I can report less than 10 people have said it didn't to me anyway ) then you will need to increase the -B number, the max is 255 which means drive APM OFF if that doesn't work then you will have to try to control the drive APM via your system BIOS or find out who makes the drive and see if they offer any utilities to control the cycling behavior.

Just remember your drive could be addressed as hda or sda - To check look in step 2. Just be sure to download the correct script and run the correct commands. While using the wrong one wont hurt anything it wont help anything either so just make sure you select the correct one and run the correct set of 4 copy commands. You just paste the commands into a terminal window ( Applications -> Accessories -> Terminal ) that saves you form having to type them :-)

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

EDIT - Locations I am recommending changed after more research to ensure fix will continue after resume if you previously used my fix then you can safely delete the script from the following 2 locations as they are not needed although if left they should do nothing to harm anything but for clean system sake I give you the following commands that if you paste whole thing in a terminal one at a time they will remove them from the unneeded locations.

These commands are potentially dangerous if altered in any way so please be careful to copy it exactly. The sudo rm command can delete any file on your Ubuntu PC and thereby render it non functional but I have written these so that you will be prompted to confirm any deletion before it occurs ( So just make sure it says the name of my script file before pressing y for yes ) and it is set to be verbose thereby telling you exactly what it is doing. I just ran both of these on my system and it did exactly what I intended. So if copied fully and not modified then these will only delete my script and nothing else and on top of that it will prompt for confirmation before deletion.

sudo rm -i -v /etc/acpi/*load-count-fix.sh
sudo rm -i -v /etc/acpi/suspend.d/*load-count-fix.sh

The only reason I use the wildcard * is so that regardless of whether you have gotten the sda fix or the hda fix it will delete either and as written the command above will delete nothing else but my scripts.
-----------------------------------------------------------------------------------------------------------------

DOWNLOAD SCRIPT "99-sda-load-count-fix.sh" IF YOUR DRIVE IS CALLED SDA BY FDISK AND THEN USE THESE COMMANDS
* make sure to set script as executable before copying or it wont work

sudo cp -f ~/Desktop/99-sda-load-count-fix.sh /etc/acpi/ac.d/
sudo cp -f ~/Desktop/99-sda-load-count-fix.sh /etc/acpi/battery.d/
sudo cp -f ~/Desktop/99-sda-load-count-fix.sh /etc/acpi/start.d
sudo cp -f ~/Desktop/99-sda-load-count-fix.sh /etc/acpi/resume.d

DOWNLOAD SCRIPT "99-hda-load-count-fix.sh" IF YOUR DRIVE IS CALLED HDA BY FDISK AND THEN USE THESE COMMANDS
* make sure to set script as executable before copying or it wont work

sudo cp -f ~/Desktop/99-hda-load-count-fix.sh /etc/acpi/ac.d/
sudo cp -f ~/Desktop/99-hda-load-count-fix.sh /etc/acpi/battery.d/
sudo cp -f ~/Desktop/99-hda-load-count-fix.sh /etc/acpi/start.d
sudo cp -f ~/Desktop/99-hda-load-count-fix.sh /etc/acpi/resume.d

This will stop all the crazy load cycles that could endanger your drives lifespan.

After you use these commands to copy it where it needs to go just REBOOT for the changes to take effect for your system.

4. CONFIRM THAT THE FIX IS WORKING AND THAT TEMP HAS NOT INCREASED TO A DANGEROUS LEVEL

Just repeat the test command from step 2 then wait 15 minutes & run it again then compare the cycle count numbers again. It shouldn't go up at all this time really but 1 or 2 would be ok ( Mine doesn't go up at all ). Also you will have this figure go up by one no matter what each time you turn the PC off and then on as the drive parks when the system is shutdown regardless of APM settings. Also compare your temperature readings now that it is not being spun down all the time.

PLEASE AT THIS POINT GO TO MY THREAD ON THIS SUBJECT AND POST WITH THE FOLLOWING INFORMATION

Model of laptop & hard disk if known.
1 hour cycle count before fix & after
Temperature before fix & after fix

THIS WILL HELP OTHER USERS OUT ALLOT AND IS ALL I ASK OF YOU

I find my temperature is fine but every situation can be different so please check yours to be sure, as the last thing I want to do is solve one problem only to cause you another.

Anything over 50C needs investigating for your equipment
And anything over 60C is critically high for sure and please PM me immediately

This is rare as I have only had one person out of 200 who have used my fix to report this and in the end he didn't need the fix in the first place as he was only getting a 12 cycle increase per hour. That said if you do get excessive heat after applying my fix then PM me and I will work to help you solve it.

From my reading 35c - 50c seem to be safe for most drives but you might investigate your drive model to be sure you are within the acceptable heat range.

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

IF YOU HAVE THE PROBLEM AND THIS DOESN'T WORK FOR YOU PLEASE LET ME KNOW BY PRIVATE MESSAGING ME, SO I CAN MAKE SURE I AM PROVIDING THE BEST SOLUTION POSSIBLE

HOW IT WORKS AND TIPS FOR IF IT WONT WORK FOR YOU

This script sets APM for your disk with a hdparm -B 200 value which is sufficient to stop all my extra cycles here anyway and should fix yours as long as you figure out what your disk device is called and get the appropriate script from the 2 below. You can use "man hdparm" without quotes to look at the manual for hdparm command and check out the -B setting for more info about what this is.

Note: Some disks are unresponsive to having their APM changed by the hdparm command, and therefore the workaround doesn't work or at least not with this value. It would be a good idea, in such cases, to disable APM in the BIOS if possible.

If this fix does not work for you ( and you tested that you have the problem ) then you will need to look in your system bios and disable APM ( Power Management - Outside of checking your drive temps the only negetive I see would be reduced laptop battery uptime as spinning the drive always will use more power. I have not however gotten anyones reports back on how much say in a % of time lost per charge this will reduce it by. If you know how long your laptop lasts on battery because you use it in that manner allot then I would be glad to hear how much this fix changed your uptime per charge as since I use mine like a desktop I don't know this myself, Please PM me with a battery uptime report anyone who can look into that.

Please make sure to remember to post your results so others can see them and the post will stay high on the lists so others can see it and be aware of this activity, thanks.

HOPE THIS HELPS YOU OUT EVERYONE - SO BE SAFE TEST YOUR LAPTOP AND USE IF NEEDED TO FIX THIS

gsiliceo
May 15th, 2008, 07:25 PM
What if the cicle count returns nothing?

nicedude
May 16th, 2008, 04:33 AM
If you mean didn't give any output and just went to a new command line, then one of 2 things. Either that you used it with device sda when your disk is addressed with hda ( check with sudo fdisk -l ) or that you didn't use sudo to give it admin power. More than likely your drive is addressed with hda though so the command for you would be the following

sudo smartctl -a /dev/hda | grep Load_Cycle_Count

Instead of

sudo smartctl -a /dev/sda | grep Load_Cycle_Count

Also make sure you installed the smart monitor tools as outlined in my directions.

gsiliceo
May 16th, 2008, 05:41 AM
No, it wasn't that, to tell you the truth my hard disk is a "special" raptor 74 gbs super fast hard disk for coporate usage,
http://www.newegg.com/product/product.aspx?item=N82E16822136033

I have another one, and thankfully the output is ok between 15 min lapse.
I wish i could diagnose the other one, are there any other tools to check SMART? or maybe i have to recompile with special flags.

nicedude
May 16th, 2008, 06:44 AM
I have a 27GB raptor in one of my PC's but I am not by that machine right now so I don't know. As far as I have seen reported this flaw mostly exists with certain laptop drives, mainly Western Digital brand has been claimed by many and I believe it is what is in my laptop ( don't want to reboot to look at bios right now since I am watching tonites episode of LOST on my laptop as I write this :-) ). So you might want to google this linux flaw and see if you can find where it affects desktops as well. I think it is rooted in a laptops aggressive power management schemes due to battery power concerns coupled with flaws in the drive firmware ( Which I found western digital has been made aware and stated they do not support end linux users, now where have i heard that before ) . I don't think there is anything special about using smart monitoring tools on your raptor disk though, did you follow my suggestion to run a sudo fdisk -l and look at your drive table and see how it is addressed? then craft the smartmon command to reflect that address. If you cant figure it out send me the output of sudo fdisk -l ( thats a lower case L not a one even though they look the same on my screen so I will point it out)
and I will customize the command for you and PM it back :-)

I am just trying to save someone a disk failure nightmare since my laptop was cycling 120 times an hour which would put it beyond manufacturers rated life expectancy in 5000 hours of use which I would like to avoid since I tend to keep PC's more than a couple years and am always squeezing life out of old ones :-)

IN MIDDLE OF WRITING THIS I CHECKED SMARTMON PROJECT AT SOUREFORGE AND IT IS SUPPORTED BY VERSION 5.37 - WHICH IS WHAT IS IN THE REPOSITORIES SO YOU SHOULD BE ABLE TO USE SMARTMON TO QUERRY YOUR DISK

So check the fdisk output or PM it to me and you should be able to run smartmon on your raptor just in case so you can sleep at ease.

over 26 hours since I applied my fix to my laptop and my count is up 8 when it would have been up 3000 or so. I can sleep with 8 much more soundly.

Now I must return to lost ep 4x12 to get my fix. PM me back to get further help so I will see it easier ( Just click my name and select send private message )

Good Luck

balagosa
May 16th, 2008, 08:28 AM
i will try that when i get home nicedude.

gsiliceo
May 16th, 2008, 08:29 AM
Oh that is not a problem i know its sda the output

Disco /dev/sda: 74.3 GB, 74355769344 bytes
255 cabezas, 63 sectores/pista, 9039 cilindros
Unidades = cilindros de 16065 * 512 = 8225280 bytes
Identificador de disco: 0x56b756b7

Disposit. Inicio Comienzo Fin Bloques Id Sistema
/dev/sda1 * 1 1912 15358108+ 7 HPFS/NTFS
/dev/sda2 1913 4462 20482875 7 HPFS/NTFS
/dev/sda3 4463 7012 20482875 af Desconocido
/dev/sda4 7013 9039 16281877+ f W95 Ext'd (LBA)
/dev/sda5 8800 9039 1927768+ 82 Linux swap / Solaris
/dev/sda6 7013 8799 14354014+ 83 Linux

Las entradas de la tabla de particiones no estn en el orden del disco

Disco /dev/sdb: 250.0 GB, 250059350016 bytes
255 cabezas, 63 sectores/pista, 30401 cilindros
Unidades = cilindros de 16065 * 512 = 8225280 bytes
Identificador de disco: 0x000c1ac5

Disposit. Inicio Comienzo Fin Bloques Id Sistema
/dev/sdb1 1 17653 141797691 b W95 FAT32
/dev/sdb2 17654 30401 102398310 83 Linux


Hehe its spanish but you know, i use the sda to hold all my OS's (xp, vista, osx, linux+swap)
I think this is a bug in the SMART monitoring tools, i blame WD, i issued the command without grep
sudo smartctl -a /dev/sda and i think this is the equivalent value

Power_Cycle_Count 0x0032 100 100 000 Old_age Always - 943
Power cycle count, would that be something different? or is the one i must monitor?

reacocard
May 16th, 2008, 08:44 AM
Does that actually work under pm-utils in hardy? It was my understanding that since pm-utils became default acpi scripts are no longer run on sleep/power events. (I have my own scripts running under pm-utils to fix this issue.)

nicedude
May 16th, 2008, 10:16 AM
To gsiliceo I don't know why smartmon tools arn't even working for you. their homepage for their project is hosted at sourceforge so you might check that out if you ever want to use them. but since Raptor drives are top of the line and not used for laptops then it is probably ok as I can only say this effects my Acer 5520-5716 laptop with Hardy Heron 32bit 8.04 for sure but the guy who first asked about it has it on his laptop as well, so if you own any laptops or manage company laptops running any linux versions I would recommend you check this out on them.

To reacocard

I can't say anything about your hardware but on my laptop with updated 32bit hardy this was happening until I implemented this script fix and I can verify 100% that this is working for me on Hardy as my load cycle count has increased only 171210 to 171218 in 30+ hours of uptime ( I have left it on constantly to check ) while before it was increasing at roughly 120 cycles per hour. I don't blame the Ubuntu makers though as this is apparently a problem with linux in general. I just wanted to save people's laptop from this danger as I realized that if their drive fries in 1 year most people would blame Ubuntu for it and get really mad, I wan't Ubuntu to grow in user base and have a great reputation so that Microsoft can lose a few of their undeserved customers (OR ARE THEY HOSTAGES) :-).

reacocard
May 16th, 2008, 10:39 AM
I can't say anything about your hardware but on my laptop with updated 32bit hardy this was happening until I implemented this script fix and I can verify 100% that this is working for me on Hardy ...

Did you install hardy via upgrade from gutsy or a clean install? If via upgrade it's possible it left acpi in place. I should test this...

EDIT: Ah, think I found why they don't on my box. acpi-support, which is what runs those scripts, isn't installed on my system becasue I installed hardy during the development cycle. However the default clean hardy install now has this package, so it should work fine for most users. Sorry about the confusion.

nicedude
May 16th, 2008, 10:57 AM
I used a clean install with the release version CD obtained from torrent download ( I love BT its my home theater ). Torrent was obtained from Ubuntu torrent site so it is all official and yet with my Acer 5520-5716 in the default APM state that hardy leaves it - this problem exists. My script ( Joke really since its just a single command ) seems to be stopping the problem from occurring though so I am happy. I just know if it effects me it effects others and so I made this thread to get other laptop users to check for this behavior on their machines. I have no idea how many other laptop brands or models it effects so I can only hope most others are safe from this . but with a potential to destroy valuable equipment it is better to be safe and check then be sorry and ordering new HDD from newegg :-)

reacocard
May 16th, 2008, 11:01 AM
I used a clean install with the release version CD obtained from torrent download ( I love BT its my home theater ). ....

I understand your reasoning and applaud it. read the edit to my post to see my resolution of the confusion.

I'm also attaching my pm-utils script here for anyone who wants it. Place it in /etc/pm/power.d/ and /etc/pm/sleep.d/ to install. Note that mine uses -B 255 unlike yours which uses -B 200, not sure how big a difference, if any, that will make.

nicedude
May 16th, 2008, 11:57 AM
You should read the man page for a this command since its use as yours is configured just completely disables power management for your hard drive with the value -B 255 , while mine with the -B 200 value stops the problem but still retains some power management for the drive it applied to. Mine gives you a little more chance of conserving power while running from battery. I would change your script to -B 200 and test that ( it can go lower 0-127 allows cycling which is what I am trying to avoid )

HERES A COMMAND FOR YOU TO TRY RUNNING WITH QUOTES AT END INCLUDED

man hdparm > ~/Desktop/"hdparm manual.txt"

It will place a text file on your desktop of the manual page for hdparm so you can learn what the commands parameter mean.

PS it works with any command that has a man page

reacocard
May 16th, 2008, 12:06 PM
You should read the man page for a this command since its use as yours is configured just completely disables power management for your hard drive with the value -B 255 , while mine with the -B 200 value stops the problem but still retains some power management for the drive it applied to. Mine gives you a little more chance of conserving power while running from battery. I would change your script to -B 200 and test that ( it can go lower 0-127 allows cycling which is what I am trying to avoid )

HERES A COMMAND FOR YOU TO TRY RUNNING WITH QUOTES AT END INCLUDED

man hdparm > ~/Desktop/"hdparm manual.txt"

It will place a text file on your desktop of the manual page for hdparm so you can learn what the commands parameter mean.

PS it works with any command that has a man page

hm, I see. I got the 255 value from the official ubuntu bug on this, but if I can get a few extra minutes by changing the value it's certainly worth looking into.

EDIT: now running using 128 instead, seems to be working just fine, not sure what impact it has on battery life yet. thanks for the tip!

Qphysics
May 16th, 2008, 04:07 PM
Maybe a stupid question. but how do I install this programm??
I cannot find it in either the programm list (or when I typ the install into a terminal.)

nicedude
May 16th, 2008, 04:35 PM
Just copy and paste the following command into a terminal and execute.

sudo apt-get install smartmontools

that will install the disk testing software which is in the Ubuntu repositories. Then look at my first post for the command to run to get the info you need to make a note of - wait 15 minutes and run the command again and note the count difference. read my first post fully to see what it should be :-)

If the install doesn't work for you with the above command then let me know and I will help you figure out why. Just click my name and select send private message and I will help you further.

Qphysics
May 16th, 2008, 05:01 PM
ok thanks.

nicedude
May 16th, 2008, 05:15 PM
Your welcome, Please post your results of your test here along with laptop model info so I can add to the list. If it isn't affecting you please post that as well so others with the same model might see.

Thanks

i speak in math
May 16th, 2008, 06:28 PM
My problem was fixed by enabling laptop-mode=true and setting laptop-mode.conf hd_power_control = 1

kuyote
May 16th, 2008, 07:04 PM
So here is the output on my 15 minute check on an asus U6s

10:45 am

193 Load_Cycle_Count 0x0032 094 094 000 Old_age Always - 13015

11:00 am

193 Load_Cycle_Count 0x0032 094 094 000 Old_age Always - 13066

So an hour would be around 200 cycles.


I'm glad I saw this before it was too late.

nicedude
May 17th, 2008, 05:20 AM
Yes kuyote that is way too high. Please post or PM me with your laptop model so I can add it to the confirmed affected list above. And please test your computer again after implementing the fix and rebooting to verify that you have it under control. At 200 per hour by my calculations that would put your hard drive past the 600,000 lifespan in 125 days or so of uptime, which is completely unacceptable by my standards.

jespdj
May 17th, 2008, 08:46 AM
The Load Cycle Count bug is old news.

Please go to this thread about it:
laptop harddrive Load_Cycle_Count issue (http://ubuntuforums.org/showthread.php?t=591503)

Note that this is not an Ubuntu-specific or even a Linux-specific bug. It also happens on Microsoft Windows on some computers.


PLEASE EVERYONE IF YOU TEST YOUR LAPTOP AND FIND THE PROBLEM IS AFFECTING YOU POST YOUR LAPTOP MODEL HERE & I WILL ADD IT TO THE CONFIRMED AFFECTED LIST ABOVE ALONG WITH YOUR NAME FOR CONFIRMING IT - THANKS EVERYONE AS TOGETHER WE CAN SUPPORT AND PROTECT EACH OTHER FROM SITUATIONS AND HIGHLIGHT THE DIFFERENCE BETWEEN LINUX AND MICROSOFT COMMUNITIES
Please DON'T SHOUT and don't panic.

Your remark about communities is nonsense and doesn't have anything to do with the harddrive issue. As if users of MS Windows don't have a strong community.

ethanay
May 17th, 2008, 08:49 AM
For those of us who have disabled advanced power management (hdparm 254 or 255) by putting hdparm scripts into /etc/acpi/*.d or /etc/pm/*.d:

install hddtemp and periodically check your disk temperature (idle and under heavy load):

sudo apt-get install hddtemp
sudo hddtemp /dev/sda
or

sudo hddtemp /dev/hda

A typical maximum is 60*C.

If your hdd is approaching this temperature, then you will need to optimize hdd i/o and take advantage of hdd spindown settings in order to keep from damaging your hdd due to heat. see this thread (http://ubuntuforums.org/showthread.php?t=772159) for more details and suggestions for taking advantage of hdd apm in order to reduce power consumption and heat

If, like me, your hdd seems to hover around 40 - 45*C then we don't have much to worry about right now, and can leisurely work our way to a more sophisticated apm state and bump protection :)

nicedude
May 17th, 2008, 10:54 AM
Thanks to ethaney for his hard disk heat check advise mine is showing 42C under heavy load with my fix enabled for my drive so I am pretty glad my temp is low for my activity and my load cycle count is virtually stopped :-)

My fix uses a value for hdparm of -B 200 which leaves some APM enabled for your drive while -B 255 is APM off for the drive and -254 is very close to off.

To anyone who wants to investigate the hdparm values start with reading the man page for hdparm as this value is explained, In general a value of 0-127 will still allow the bad behavior for sure since in that range it allows for load cycling. some other posts on these forums and else where call for varying values to stop this behavior including 255 to just turn off drive APM but since in my tests -200 works and seems reasonable as it just gives me a chance at some power management with my drive. If anyone actually tests different parameters and would like to share their findings then by all means I would welcome hearing from you.

I just want all laptop users to test for this as it is obviously occurring and other than the test with smart monitor tools I would have never known until too late.

My count increase now stands at 13 new cycles in 3 days of up time compared to the 8600 it would have been before I patched it up. Since my disk has a rated life expectancy of 600,000 cycles I don't think 3 days up time should equal 1.5% of my drives load cycle life expectancy.

I see someone thinks this is an old problem well great and just like cancer its still here apparently and I along with others ( With years of Ubuntu under their belt ) had no idea. So if you already know or have fixed it for your self some other way, awesome I am happy for you. But to anyone that has not and had no idea about it then please check your laptop out for yourself and if affected fix however you want, or don't fix it at all if it doesn't bother you.

In future I might try to use powertop and compare some power drains with different parameters if I get a chance and report them here. If anyone else wants to do that then be my guest and have at it, I am sure someone that reads this can decipher power draws on the system for varying values much better than I.

Good luck to all with whatever you do :-)

FokkerCharlie
May 17th, 2008, 11:41 AM
OK, thanks for this.

My Acer Aspire 5920 (120GB version) does not seem to be affected:


193 Load_Cycle_Count 0x0012 100 100 000 Old_age Always - 1348

15 mins later:


193 Load_Cycle_Count 0x0012 100 100 000 Old_age Always - 1348

Cheers
Charlie

issashu
May 17th, 2008, 12:49 PM
Not at my home laptop right now, but will check once I get home.
Right now I can say the following concerning the vaio VGN FE-21m:

There was a problem regarding this model and it's hard disk for quite a while now. There is a clicking sound coming from the hdd after a bit of work. The sound can be heard on regular intervals of around 10 s. It has been said thids is because of the hdd heads parcking themselfs, but this happens even while activly working with the hdd (copying files, playing games, etc). This lead to a lot of returned laptops for service with broken HDDs...

Now after reading the thread, I think this is similar problem, so all vaio fe-21m users take a good check of your situation.

Will post exact hdd model and results once I get home.
The hdd was a fujitsu model (can't remember which one).

Anduu
May 17th, 2008, 07:29 PM
I can confirm Acer Aspire 7520-5839 with an ATA Hitachi HTS54251 is effected.

Implemented your fix and all is well now.

tezer
May 17th, 2008, 08:26 PM
Hi!
My FUJITSU MHV2080A (ASUS A6F) doesn't seem to have the SMART system:


smartctl version 5.36 [i686-pc-linux-gnu] Copyright (C) 2002-6 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

Device: ATA FUJITSU MHV2080A Version: 0000
Serial number: NT9WT6826F0B
Device type: disk
Local Time is: Sat May 17 20:18:17 2008 BST
Device does not support SMART

Error Counter logging not supported

[GLTSD (Global Logging Target Save Disable) set. Enable Save with '-S on']
Device does not support Self Test logging

The temperature is 43 - 44 without any intensive disk activity.
Anything else I can do to ensure my HDD isn't going to leave the world soon?

ASULutzy
May 17th, 2008, 10:35 PM
HP dv6000 is affected. Got around 14 cycles in 10 minutes, going to go ahead and check the drive temp now to make sure that's still ok after applying the fix.

ASULutzy
May 17th, 2008, 10:40 PM
My hard drive temperature after implementing that fix is around 50C under normal usage, I need to check heavy usage... That seems a little warm

johnl
May 17th, 2008, 11:06 PM
My Gateway M-1617 seems to be affected. Count went from 27741 to 27747 in about 5 minutes.

The drive is a Western Digital: ATA WDC WD2500BEVS-2

Ripfox
May 17th, 2008, 11:12 PM
My hard drive temperature after implementing that fix is around 50C under normal usage, I need to check heavy usage... That seems a little warm

That's not warm. I had this laptop and it's normal operating temperature is 50-55c.

Please take down the panic level. If you want to help, everyone can read fine in normal text, you come off as shouting and panicked. Creating panic never leads to good results.

miga
May 17th, 2008, 11:54 PM
guess it's worth mentioning, approx 10 minute wait gave me an increase of 8 cycles..

on a Lenovo thinkpad R61i, using a
120Gb drive
Device Model: FUJITSU MHY2120BH
Serial Number: K434T812591M
Firmware Version: 0084000D

Arthur Archnix
May 18th, 2008, 12:43 AM
here's the relevant launchpad bug: https://bugs.launchpad.net/ubuntu/+s...ort/+bug/59695

Your link goes nowhere for me.


The problem occurs when APM ( power management ) is constantly parking and unparking your HDD and is a flaw with only certain brands of drives ( Western Digital for sure and they have no plans to fix the firmware as they stated they do not support end linux users ).


That's a horrible summary of the problem.


Someone else ( Kestol ) mentioned below asked the question about this abnormality and at first I thought he was dumb ( Sorry Kestol ) then I found that my laptop suffered from it so I made this thread to WARN YOU to test yours as well.


There are plenty of these threads out there, one in particular has been around a long time, and has much better information.This one by Ubuntu Demon for example...

Link (http://ubuntuforums.org/showthread.php?t=591503)

Ubuntu demon has been on this issue for a long time. I'd advise you to read his thread and make sure your own is in line with his information. If you differ or disagree, you should explain why.


First thanks to Kestol for figuring out this was taking place.

The bug was first reported by Gilles Schintgen in 2006. Here's a link to the bug:

Link to original bug (https://bugs.launchpad.net/ubuntu/+source/acpi-support/+bug/59695)


But for people who would rather protect themselves than role the dice with their stuff, I have made a simple script to fix this problem.



Reasonable Limits / Criteria for a fix:
* There should be fewer than ~15 load cycles per hour, except during heavy usage while on battery.
* This provides a life expectancy of over four years, which is reasonable for a hard disk.



I have found that a setting of -B 200 works just fine here and my cycle count has gone up 8 in 12.5 hours vs 120 per hour prior to this fix, so I am happy as a pig in slop.


This is not good info. I'd advise people against using this guys fix. Read the bug. Read Ubuntu Demons thread. At a minimum, this guy should have included a warning that disabling power management can lead to an increase in HD temperature. Which could also decrease the life of your hard-drive.

His script is not in line with the bugsquads temporary workaround. And it is not in line with what Ubuntu Demon has written about it. Both links are provided for you. Working links.

This guy is Johnny come lately to the issue, and while I applaud his enthusiam and desire to help others, he's in over his head and people who read this thread ought to know that there are better sources of information out there about this bug.

Start with demon's post. He's got lots of great links for you to learn more. One more time, here it is: Demon's discussion of Laptop Load Cycle issue (http://ubuntuforums.org/showthread.php?t=591503).

Anduu
May 18th, 2008, 01:08 AM
If you will notice that thread is archived and has seen no activity since October of last year.{Edit...}Oops my browser wouldn't load past the first page...anyway not everyone knows to check archeived posts.

I applaud the OP for bringing this issue to light again because as I said I had completely forgotten about the issue and just recently purchased my first laptop and guess what...My drive was excessively cycling.Who knows what effect this would have had on the life of my drive.

KOTAPAKA
May 18th, 2008, 02:15 AM
If this fix does not work for you ( and you tested that you have the problem ) then you will need to look in your system bios and disable APM ( Power Management )
this shouldn't be a problem for a desktop system - but would affect laptop battery life

I never use my laptop on battery (well I do once, twice a year). So is disabling my APM going to help anything? I am not sure my HDD is affected but as I never use the battery...
And what does this mean (output of smartctl -a /dev/sda2 | grep Load_Cycle_Count )

193 Load_Cycle_Count 0x0032 096 096 000 Old_age Always - 85070

ethanay
May 18th, 2008, 03:58 AM
I agree that the information isn't all correct and top-notch, but
1) neither was UbuntuDemon's when it first came it (process of discovery...)
2) regardless it is no reason to bite the hand of someone offering a gesture of good will

nicedude
May 18th, 2008, 06:40 AM
My comments deleted here as they were made in haste in response to a PM and I now retract them.

Please everyone I make no claims of being an expert in this situation or Linux in general I am just trying to help others, nothing more but also nothing less. So anyone with more knowledge than I please PM me and I will listen to what you have to say on this topic and revise my original post & script fix accordingly.

To any that give help in this matter I thank you as I feel like a one man band right about now.

Ripfox
May 18th, 2008, 07:23 AM
Gee whiz I just started this thread to try and warn people that this was occuring as I didn't see it in any of my thread reading or mentioned by anyone. If you don't like the advice then don't use it but there is no harm in installing the smart monitor tool package and checking what your situation is.

As far as being a johnny come lately since I saw nothing about this until someone asked in the absolute newbie section I was not aware of their being posts by johnny came early, and since its not like I am being paid per click or something to warn anyone I am doing it out of the goodness of my heart.

To Arthur above please feel free to make your own thread on this issue and I can stop having to post on this one. Or else go take a long walk on a short pier there in Finland, as I don't see you trying to contribute anything but smart mouth gibber jabber.

And your advise about my -B 200 value while stating you saw another suggesting -254 well -200 lets you have more APM management not less so try reading the man page for a command before opening your mouth "man hdparm" if you need help with that. Actually you can use any -B value you want between 128-255 and it should stop this but I chose 200 as a safe value with a little more chance of APM then 254 or 255 which just turns off APM completely for your laptops hard drive. So if you use 254 then 200 would only allow more chance of APM kicking in not less and is not dangerous for anyones system. I actually read the damn man page did you?

Sorry for the anger but this guy has my blood pressure up as I am just trying to help out newbies to linux (like myself) and had no idea that there was already another thread anywhere on these forums about this. I am truly sorry I tried to help now as if ridicule and smart mouth is all the reward I get for trying to help then I don't want it.

This is my final post in this thread and I will just let it be what it is. Arthur Archnix can help you all as he seems to be far more knowledgeable in his mind than I, so feel free to PM him for assistance in this matter as I am done with this thread and am going to delete my subscription to it. I am sorry that I tried to help people & I should be ashamed of myself for it.

To anyone I did help you are welcome

No offense, but this is exactly what is wrong with your posts. You sound like you just drank 50 red bulls and then posted a loud "warning" in almost all capital letters. Did you search the forums for this problem before you posted this thread? Archives are still searchable, and yes, this has been around for a long time and there is a fix in place. It wasn't you trying to help that was the problem, it was the lack of research and the panicked tone you took on. Ease into things a little more and you will get better results. I'm sorry If that hurts your feelings to the point where you proclaim you are "sorry you tried to help anyone"

beyboo
May 18th, 2008, 07:56 AM
I confirm that my 2 months old Acer Aspire 5920-5A2G16Mi-ZD1 has the bug. It rises by an alarming 30-40 units in appx. 15 minute intervals.

Hard Disk Details

Device Model: Hitachi HTS542516K9SA00
Serial Number: 080202BB2C00WGK1WE8A
Firmware Version: BBCOC31P
User Capacity: 160,041,885,696 bytes


I am in India and this acer model 5920-ZD1 is a regional model and maybe different for the Europe / US Market. Check out the specs here.

http://www.acer.co.in/products/product_explore_view.asp?pid=89&model=Aspire%205920%20&cid=5#specifications

I will implement the script you got and report back if I find a difference.

Thanks Nice_dude !!

Edit :

This from nice_dude was key to alerting me on my laptop. Further reading as recommended by some others (Arthur) led me to using the script which conditionally tweaks the load/unload using hddparm. However my first thanks to nice_dude for alerting me on this one. Thanks to the splendid and mind blowing ubuntu community who keep such a close watch on things and come out with discussions, flares, fights and ultimately great support for end users !!!

Arthur Archnix
May 18th, 2008, 10:19 AM
I haven't attacked NiceDude's intentions. In my original post I believe I even applauded his desire to help others. What I have done is tell him and nice dude there are better sources of information on this problem. IMHO, there are better scripts as well. Ubuntu demon runs a check to see whether you're no battery or not. If you're on battery it sets your HDparm value lower, since you want head parking to protect from bumps. When it's plugged in it sets it higher.

I'm trying to help people too Nice Guy. If I could warn them that your information and workaround may not be the best without also implying that yours isn't very good I would. Try to keep in mind that my goal isn't to hurt your feelings, it's to make sure that people here have the best information.

I haven't upgraded to Hardy yet, but last time I checked this issue had been fixed upstream in debian and the accepted fix was to enable laptop mode.

And no, in response to your PM I am not going to summarize the problem and PM it to you. Read the threads yourself. Why don't you start by comparing your script with Demons? All you do is set HDParm to 200. It should be lower when on battery. Telling people that 200 is a safe option when on battery is not a good idea. Someone's hard-drive could die from a bump because the head didn't park.

Finally, I notice that you didn't bother to edit your original post and include a warning about the danger of overheating your hard-drive. Apparently you'd rather defend your honor and act hurt than actually be a nice dude and continue to update this post:

To everyone else, this fix is not to be used lightly. It sets one hdparm value. You may as well just put the hdparm command in rc.local and forget about the script if all you care about is setting it once and forgetting it. Demon's does a sanity check. If you're on batteries is sets a lower value to protect your hard-drive. If you're on ac power it sets it higher. If like NiceDude you think 200 is a better value, then change that. It's easy to do. And keep an eye on your hd temps. And if you're on Hardy... try enabling Laptop mode first. See if that fixes the problem

beyboo
May 18th, 2008, 10:39 AM
To everyone else, this fix is not to be used lightly. It sets one hdparm value. You may as well just put the hdparm command in rc.local and forget about the script if all you care about is setting it once and forgetting it. Demon's does a sanity check. If you're on batteries is sets a lower value to protect your hard-drive. If you're on ac power it sets it higher. If like NiceDude you think 200 is a better value, then change that. It's easy to do. And keep an eye on your hd temps. And if you're on Hardy... try enabling Laptop mode first. See if that fixes the problem

I acknowledge and support you on this. I read up the "original" post on this issue and used the script which sets it to disabled on AC and to 128 on battery. I agree its much safer.

I guess we require both nice_dude's to alert users and ppl like you to moderate and keep watch so that ppl like me benefit

I think both of you are essential parts of the system. As I see it, if this wasnt seen by me, I wouldnt have landed in to the refined solution at all !!

:KS

halfmanhalfbug
May 18th, 2008, 11:18 AM
Yup, I have the bug: 24 cycles in 15 minutes on a clean installation of Hardy (The bug's not fixed yet, even in Hardy, even after 2 years!)
Model Number: Hitachi HTS541616J9AT00
Firmware Revision: SB4OA70H
This explains the premature death of my first drive. My current drive is only about 6 months old and already has 639265 load cycles. Unbelievable.

Arthur Archnix
May 18th, 2008, 12:11 PM
By the way, nice Dude, I think your documenting of the hard-drives affected and models affected is outstanding. A long-overdue contribution that adds real value to subject. I hope you continue to update and refine this thread with all that you learn.

I have an HP 530 and its affected, so are most people I've chatted with on the HP 530 support thread.

Crackh34d
May 18th, 2008, 12:35 PM
I have a 8 months-old Lenove T60p. My cycle count is already 793000 and counting rapidly. Your fix did work for me, thanks

volkswagner
May 18th, 2008, 01:34 PM
Thanks nicedude,

I remember reading about this last year when I fist started with Ubuntu. I did not see any solutions.

I have Toshiba Satellite A135-S2246. My Load cycle count was running about 100/hr. I have a count of 130467, after one year of moderate usage on this new machine. I rarely leave the laptop on overnight.

After applying your script, my count has not increased after 12hrs running. :)

I will post back with the hard drive specs, from BIOS.

Here is what I find in system information.
-SCSI Disks-
ATA TOSHIBA MK8032GS

nicedude
May 18th, 2008, 02:18 PM
Actually I changed my mind on the abandoning contributing to this thread as I don't think I should let anyone chase me away from helping people. Arthur please look at my edited first post as there are now 3 working links to offsite references to this problem and my scripts values are within acceptable ranges of fixing this and they are at the lowest APM stopping end of the range to allow the most chance for APM to have a positive effect for users battery life. I actually read the hdparm man page and spent several hours researching this problem to come up with that value and over 10 people have PMed me to say thanks and that my script fixes the issue for them. You just really got my blood boiling that you would spend that much time to critique my post and call me a johnny come lately ( Thats just plain rude ) when you didn't offer any corrections or to try to help. If you Arthur ( Or anyone for that matter ) can figure out a better -B hdparm value (preferably that still leaves the possibility of power savings as I could have just used 255 and turned it off ) or any other way to insure a fix that will work for everyone I will be glad to edit my first post and not only use your value or method but also give you credit for it. I just don't appreciate scorn for help as I have made this thread out of pure concern for my fellow human beings as I know if Linux tears up there equipment they will go back to Microscum Winblows and talk trash about linux as well, And as I have ditched MS OS for good as of 2 months ago for this great OS that is Ubuntu I will try to give as much help as I can to this community since the OS is free and I am benefiting from it by using it on 5 PC's for free. While I am a linux newbie I have been fixing & tweaking PC's since DOS 5.0 so I feel I do have some PC experience under my belt to justify my attempts at intelligent help. So everyone please keep sending me your confirmed affected and confirmed unaffected reports and I will continue to add them to the first post as they are received. Please PM them to me though so I don't have to read all the posts here to see them as I might miss them.

To Arthur I am just gonna wipe your slate clean in my mind and start over with you since you did make a somewhat apology above and if you would like to contribute to this topic please PM me with any relevant results and I will add them where relevant and give you the credit for the work.

And I apologize Arthur for the step off a pier remark in my earlier post and will amend it here to read take a short walk on a long pier in Finland and enjoy the sunset :-)

Arthur Archnix
May 18th, 2008, 02:29 PM
You're right NiceDude. The Johnny Come Lately remark was rude. I apologize for that.

RRFarFar
May 18th, 2008, 02:37 PM
I confirm that such problem exist on Dell Inspiron 640M/E1405. So I followed the instructions and the number keeps changing (I have made chmod +x command for the script in terminal):

$[~] 16:00 rfar$ sudo hddtemp /dev/sda
/dev/sda: WDC WD1200BEVS-75RST0: 42C
$[~] 16:00 rfar$ sudo smartctl -a /dev/sda | grep Load_Cycle_Count
193 Load_Cycle_Count 0x0032 189 189 000 Old_age Always - 34281
$[~] 16:00 rfar$ sudo smartctl -a /dev/sda | grep Load_Cycle_Count
193 Load_Cycle_Count 0x0032 189 189 000 Old_age Always - 34281
$[~] 16:00 rfar$ sudo smartctl -a /dev/sda | grep Load_Cycle_Count
193 Load_Cycle_Count 0x0032 189 189 000 Old_age Always - 34282
$[~] 16:06 rfar$ sudo smartctl -a /dev/sda | grep Load_Cycle_Count
193 Load_Cycle_Count 0x0032 189 189 000 Old_age Always - 34282
$[~] 16:06 rfar$ sudo smartctl -a /dev/sda | grep Load_Cycle_Count
193 Load_Cycle_Count 0x0032 189 189 000 Old_age Always - 34282
$[~] 16:11 rfar$ sudo hddtemp /dev/sda
/dev/sda: WDC WD1200BEVS-75RST0: 43C
$[~] 16:11 rfar$ sudo smartctl -a /dev/sda | grep Load_Cycle_Count
193 Load_Cycle_Count 0x0032 189 189 000 Old_age Always - 34282
$[~] 16:11 rfar$ sudo smartctl -a /dev/sda | grep Load_Cycle_Count
193 Load_Cycle_Count 0x0032 189 189 000 Old_age Always - 34291
$[~] 16:31 rfar$

The temperature is OK, but the number is changing
Could you advise something please???
Also I have another drive - USB Western Digital Passport. Does this bug effects USB? A newbie question.
Maybe it is easier to disable laptop mode. What will be the consequences of that????
Cheers for your help

nicedude
May 18th, 2008, 03:24 PM
Some have implied the laptop mode tool is supposed to fix this but I did not have that result here. If someone with uber linux knowledge or knowledge about the results of hdparm could PM me with info about hdparm -B values to fix this situation I would be glad to receive and then implement such information into my script fix and directions. I see some say a on battery value of 128 would be better for allowing parks, I will look into that and test it to see if it also resolves the issue for me here on my Acer 5520 if so I will update my original post and implement a 2 script per machine solution for different modes of operation or just use the 128 if it seems to keep my parks low while on AC power. The value I see recommended on many other sites is either 254 or 255 and either of those would leave a hard disk even more susceptible to being damaged while on the go than 200 would so I felt it was/is safe to use 200 as the value. That said once again if anyone has a intimate knowledge with the hdparm command please PM me advise on values as my only intent is to save someone from a bad situation, Not to cause one.

To all that PM me models affected or unaffected or give knowledgeable advise I thank you very much.

PS after I post this I am just going to edit my first post with instructions for how to check your disk temps and give the credit for that to ethanay as he deserves it for pointing that out first ( Along with directions on how to do it :-) )

nicedude

dmcn
May 18th, 2008, 03:52 PM
I have a machine built on a SAMSUNG HM121HC which is affected. My setup is different, though - it's the system disk in my file server (running Debian) in my home office, connected via a converter to the onboard IDE-port on a Mini ITX motherboard. I guess some kind of disk power management is doing something bad. I'll try to boot into Samsung's health tool to see if I can turn it on, but for now, Samsung HM121HC is confirmed.

nicedude
May 18th, 2008, 05:12 PM
I have yet to hear from anyone with inflated temperatures for their disk after implementing my fix and mine is still showing 42C over 2 days uptime strait under heavy loads and while idle. So if anyone gets a higher temp due to my fix or any other fix for this problem I would really appreciate the information so I can help people the best I know how with this issue. Thanks in advance for any info.

boppp
May 18th, 2008, 05:17 PM
I have applied your fix on my Asus F3JA (which had the problem, but already reported that on PM).

Just unrarred a 5 gb file and my hdd temp is still 42 degrees celcius. So no problem there :)

Anduu
May 18th, 2008, 05:47 PM
I have yet to hear from anyone with inflated temperatures for their disk after implementing my fix and mine is still showing 42C over 2 days uptime strait under heavy loads and while idle. So if anyone gets a higher temp due to my fix or any other fix for this problem I would really appreciate the information so I can help people the best I know how with this issue. Thanks in advance for any info.

No issues with temps so far...max I have seen is 40C

yellowdude
May 18th, 2008, 05:47 PM
Hey Mod Should we make this into a Sticky?

Valloric
May 18th, 2008, 07:41 PM
I can confirm that the laptop HP nx6125 has this issue, and I am VERY VERY grateful to nicedude for making this thread because I didn't catch the older ones.

I ended up using the solution from opensuse that Arthur posted (thanks to you too dude) because it looked more robust (again, it can be found here (http://en.opensuse.org/Disk_Power_Management)).

My cycle count (without the fix) went up by 32 points in 15 mins. Now, after the fix, it has been 20 mins and not a single point increase. Also, temp is steady at 40 C.

Again, a big THANK YOU to nicedude, I caught this thread by accident.

RRFarFar
May 18th, 2008, 10:18 PM
To the NICEDUDE (the one who created this post)
I have already thanked you for your post. With your help and help of some other posts I have managed to solve my issue with Dell 640M. It is all right now.
Anyway this issue is quite serious, at least in my opinion. Can you or anyone tell if there are some other serious issues with Ubuntu (hardware or security)???? I know that this hardrive issue is not really Ubuntu or even Linux issue, but anyway there can be some other issue. For example I cannot burn any data cd/dvd on my machine. The only thing I managed to burn was an iso image.
Once again I think issue my be solved, and probably most of the new Ubuntu users do not even know about this issue. i am very sorry for them, since there hardrive might now live till version 8.10
Thanks again to everybody

nutpants
May 19th, 2008, 02:45 AM
in ten minutes on a dell xps1210

193 Load_Cycle_Count 0x0022 090 090 000 Old_age Always - 20822

193 Load_Cycle_Count 0x0022 090 090 000 Old_age Always - 20863


thanks for the tip


nutz

nicedude
May 19th, 2008, 04:06 AM
To RRFarFar and anyone else concerned about safety & security ( thats all of us though right? ) this is the first dangerous issue I have seen or heard of. In general Linux ( Ubuntu included ) is more secure than Windows, Bill Gates would claim otherwise but consider the source :-).

As far as security in general on your Linux box you are immune at this time from viruses, there are proof of concept Linux viruses but none are "in the wild" as they say. So installing software such as clamwin from the repositories is only for scanning Windows partitions and files for Windows based viruses. This is a powerful way to find nasty windows viruses by the way since it is one way to be sure that they are not manipulating the output of a virus scan ( to hide themselves ) such as is possible for some viruses to do in windows.

As for a firewall you have one that comes built in to Ubuntu called IPtables but it is configured from the command line so if you want an easy way to configure it using a GUI you can use either "firestarter" or "guarddog" as they are in the repositories for you. I prefer guard dog and found it to be simple and yet powerful in operation.

The biggest threat I would probably have to say would be from a rootkit as they do exist in Linux just as they do in windows. For checking for those you can get "ckroot" and or "rkhunter" both in the repositories and while rkhunter is fairly strait forward to use, the chkrootkit is way more complicated so I would recommend for an advanced user only. The number one way to avoid being affected with a rootkit though is to never install software that is not either from the repositories or from a trusted source and even then you can check the MD5 hash vs the correct one on the authors webpage etc. to verify that the version you got hasn't been tampered with.

In short if you are concerned or keep valuable data on your machine I would recommend you get Guard Dog & Rkhunter. then if you are an advanced user and want to learn before you use then get Chkrootkit and read up on it as its not as simple as rkhunter.

Command to install both guarddog & rkhunter at once
sudo apt-get install guarddog rkhunter

Guard dog will be under Applications -> Internet, as for rkhunter you just type the commands in a terminal to run it, common commands below

RKHUNTER COMMANDS
Command to update definition files
sudo rkhunter --update
Command to perform a check of the system for rootkits
sudo rkhunter --check

I hope this alleviates some of your fears and if any Linux gurus here know of anything else I didn't mention please chime in as I would like to know as well.

Nicedude

advark
May 19th, 2008, 04:16 AM
Hello after using the command
sudo smartctl -a /dev/sda | grep Load_Cycle_Count
in an interval of 6 hours give me change of cycles of 9
9 is a good change of cycle count in 6 hours?

The reason of my question is if necessary for me to use your script Nicedude.


Regards

PS: My laptop is a sony vaio vgn-n350fe

ASULutzy
May 19th, 2008, 04:22 AM
At 200 my drive was still cycling way way more than I was comfortable with. I've set mine to 255 now and I'm going to check the temperatures of the drive while archiving and then extracting large files.

I never really have my laptop unplugged, so battery life is not a real concern for me.

Out of curiosity, what is the default value of hdparm? Just in case I ever do unplug it, I'd probably like to reenable APM power management on the drive then

reacocard
May 19th, 2008, 04:23 AM
Hello after using the command
sudo smartctl -a /dev/sda | grep Load_Cycle_Count
in an interval of 6 hours give me change of cycles of 9
9 is a good change of cycle count in 6 hours?

The reason of my question is if necessary for me to use your script Nicedude.


Regards

PS: My laptop is a sony vaio vgn-n350fe

9 is a very good value for 6 hours, you have nothing to worry about.

Anduu
May 19th, 2008, 04:26 AM
Hello after using the command
sudo smartctl -a /dev/sda | grep Load_Cycle_Count
in an interval of 6 hours give me change of cycles of 9
9 is a good change of cycle count in 6 hours?

The reason of my question is if necessary for me to use your script Nicedude.


Regards

PS: My laptop is a sony vaio vgn-n350fe

According to nicedudes criteria less than 15 cycles per hour is acceptable.

ASULutzy
May 19th, 2008, 04:33 AM
Oh, also, I did the math, I've been averaging about 50 cycles per hour (This is having never used battery power without being plugged in)

Considering drives are supposed to last for 600,000 cycles, and considering I've been averaging 50 cycles per hour, that's a mere 12,000 hours of expected life, which when converted to years is only 1.37 years.

Yea, I'm leaving mine at 255 unless the temps get too high.

Now I'm actually concerned about booting into Windows and if Vista will cycle my hard drive too much, is there any concern with Vista doing this?

Anduu
May 19th, 2008, 04:40 AM
Oh, also, I did the math, I've been averaging about 50 cycles per hour (This is having never used battery power without being plugged in)

Considering drives are supposed to last for 600,000 cycles, and considering I've been averaging 50 cycles per hour, that's a mere 12,000 hours of expected life, which when converted to years is only 1.37 years.

Yea, I'm leaving mine at 255 unless the temps get too high.

Now I'm actually concerned about booting into Windows and if Vista will cycle my hard drive too much, is there any concern with Vista doing this?

You leave your laptop on 24/7? Yikes!

If you factor in 8 hours per day you can multiply your figure by 3 giving you over 4 years ;)

ASULutzy
May 19th, 2008, 04:48 AM
Heh, I come close to 24/7.

I use it at work all day, and when I come home I'm usually playing around on it, and at night I hook it up to the TV and use Windows Media Center to watch some shows while falling asleep. There's probably 3 hours a day or so where it's off. (Oh, speaking of which, if anyone knows how the heck it's possible to get S-video working in Ubuntu with an Intel i965 on board GPU, I'd love to hear it since that's the only reason I even have Windows still installed on this machine)

But yea, drive temps are always <50C even with APM off, so I'll be leaving that value at 255.

nicedude
May 19th, 2008, 05:54 AM
No it is not necessary for you advark as your figure of 9 per six hours works out to 36 per full day and that equates to 13140 per year of so your hard drive will not die from this bug in 20 years of uptime so it will no doubt ( I hope long from now ) die by some other natural death after a long time of service but certainly before 20 years so you are in the clear my friend and I will post your model as being reported OK.

Thanks for the info on your model and I am glad to hear you don't suffer from this :-)

To all here the 15 per hour is "ok" is not really my criteria it is from the bug tracker launchpad article on this topic so everyone please just do the math for your situation based on hours per day of operation and then decide if how long your disk will take to reach the 600,000 barrier is acceptable to you or not. I will calculate some examples for everyone here now.

Note some peoples drives could be rated to last 1,000,000 cycles but since I have no idea what ones are I stuck to the lower number for safety sake.

If you have your laptop on 8 hours per day
15 cycles per hour - how long till 600,000 barrier 13.5 yrs
25 cycles per hour - how long till 600,000 barrier 8.1 yrs
50 cycles per hour - how long till 600,000 barrier 3.9 yrs
75 cycles per hour - how long till 600,000 barrier 2.7 yrs
120 cycles per hour - how long till 600,000 barrier 1.6 yrs
200 cycles per hour - how long till 600,000 barrier 1.0 yrs

If you use it 12 hours per day
15 cycles per hour - how long till 600,000 barrier 9.0 yrs
25 cycles per hour - how long till 600,000 barrier 5.4 yrs
50 cycles per hour - how long till 600,000 barrier 2.6 yrs
75 cycles per hour - how long till 600,000 barrier 1.8 yrs
120 cycles per hour - how long till 600,000 barrier 1.1 yrs
200 cycles per hour - how long till 600,000 barrier 8.0 months

If you leave it on almost 24/7
15 cycles per hour - how long till 600,000 barrier 4.5 yrs
25 cycles per hour - how long till 600,000 barrier 2.7 yrs
50 cycles per hour - how long till 600,000 barrier 1.3 yrs
75 cycles per hour - how long till 600,000 barrier 10.9 months
120 cycles per hour - how long till 600,000 barrier 6.8 months
200 cycles per hour - how long till 600,000 barrier 4.0 months

More than 200 cycles per hour and you better fix it right now as that would be really bad and worse than mine was :-)

Some of the above are sorta rounded off so I could hurry and do it but they are all real close :-)

Zip247
May 19th, 2008, 07:25 AM
I tested my Toshiba Satellite A75 S211 and the number went up 30 in 15 minutes. That does not sound too good.

Downloaded your file, set it as executable and copied it to the needed places and rebooted.

New 15 minute count is 3.

My hdd temp was 32 prior to the fix now its 35, I can live with this, Its still on the low end.

Thanks for this post. I may even play with the setting to get it lower.

russo.mic
May 19th, 2008, 09:43 AM
The Load Cycle Count bug is old news.

Please go to this thread about it:
laptop harddrive Load_Cycle_Count issue (http://ubuntuforums.org/showthread.php?t=591503)

Note that this is not an Ubuntu-specific or even a Linux-specific bug. It also happens on Microsoft Windows on some computers.


Please DON'T SHOUT and don't panic.

Your remark about communities is nonsense and doesn't have anything to do with the harddrive issue. As if users of MS Windows don't have a strong community.

Your critizing him for pointing out a problem with an OS on the forums dedicated to that OS and offering a solution...is that correct?

Just wanted to make sure I understand fully. Why don't you check out this link:

http://en.wikipedia.org/wiki/Ubuntu_(philosophy))

Nicedude, I just tested my Apple Macbook Pro, which has the FUJITSU MHW2120BH hardrive in it, which is in your list. I don't seem to have this problem, as my load_cycle_count hasn't increased one tic in almost 30 min. This could be due to apple power managment. Just letting you know!

Russo

llevering
May 19th, 2008, 10:39 AM
You can add the Dell Latitude D520 as being affected too. Or at least the ones with 120gb Harddisk as they have a Toshiba 1234GXS harddisk.

halfmanhalfbug
May 19th, 2008, 10:43 AM
Hi nice dude. Thanks for all your work. I have a dual boot system (Hitachi HTS541616J9AT00, HP L2000) so I tested Windows XP. It only made 2 cycles in 1.5 hr (and the cycles may have been the two power-offs) compared to 96 per hour in Ubuntu. The developers over at launchpad seem to regard this problem as the HD firmware or the BIOS and they still have not even decided how to begin making the fix after two years of knowing the bug exists. Let's face it, if Windows can handle it and Ubuntu can't, then Ubuntu has a MAJOR bug. I can't really begin to describe how disappointed I am in Ubuntu/Linux for this.

KOTAPAKA
May 19th, 2008, 11:10 AM
I just found out that by changing the hdparm.conf the changes do not take effect. Anyone got any solution?

Kestol
May 19th, 2008, 01:23 PM
Well... Finally i decided to post in here.. So ye i was affected and i'm gald nicedude made an awesome fix:)
Better then the one by jakon since it still enables APM (No offense Jakon:))

Well.. Here are my infos...

Laptop: Packardbell EASYNOTE_MX67-O-070 (More info can be found here: http://support.packardbell.com/uk/item/?m=home&sn=103136450326)
And i was getting an increase of about 35 parkings an hour...:)

As for the fix u gave.. I checked my heating and everything seems to be as usual:) SO no heating probs here:)

chriswyatt
May 19th, 2008, 02:43 PM
DV2058ea, tried it again after about 4 minutes and it went up by 4, I think my laptop has the bug (did this when running in battery mode with laptop mode enabled). I can hear my hard drive constantly spinning up, can't be good. Gonna try this fix now.

RRFarFar
May 19th, 2008, 04:35 PM
I was already here, but forgot to ask.
Does anyone have a clue whether this bug affects USB external harddrives???I have one and its Western Digital)))
Those terminal commands above do not work for usb drive, even if I change sda to sdb
Thanks in advance

G@B0
May 19th, 2008, 04:43 PM
Add to the list.
Sony Vaio VGN-FZ180E

Drive: TOSHIBA MK2035GS

It was at 80/hour before the fix. Now it's working great.

The temperature on this laptop is awful. It's always more than 50.
It goes from 60-80 while using Flash.

reacocard
May 19th, 2008, 05:05 PM
I was already here, but forgot to ask.
Does anyone have a clue whether this bug affects USB external harddrives???I have one and its Western Digital)))
Those terminal commands above do not work for usb drive, even if I change sda to sdb
Thanks in advance

USB drives are not affected. The chip in the drive takes care of all power management stuff so the manufacturer can make sure that it works properly. The OS is not involved.

Luke has no name
May 19th, 2008, 05:15 PM
Why isn't this Sticky'd?

kuyote
May 19th, 2008, 05:53 PM
I just tested my laptop again, after using it all weekend, and can say that the fix has definitely slowed down my cycles.

Model: Asus U6s
Disk: Seagate ST9160821AS

15 minute average before: 50 cycles
last 2 days: 200 cycles

darco
May 19th, 2008, 06:21 PM
I confirm I have this bug on my Toshiba Satellite A105-S4254.
After running the Load cycle app, I was getting 41 cycles/15mins.

Downloaded fix and now after 15mins, no cycle count increase.

Downloaded and ran hddtmp and I show 33 C

thanks NiceDude!

good luck

darco

pharnet
May 19th, 2008, 08:40 PM
Thanks for the solution, please add mine to the list:

Acer Aspire 5570-4315 & Hitachi HTS54161

joshoekstra
May 19th, 2008, 10:44 PM
And mine: HP 6710B with TOSHIBA MK2546GSX

Thanx for fix, anyway to fix the old-age label? :-k

linuxuser187
May 19th, 2008, 10:47 PM
Thank you! worked for me too, i was getting 17 cycles within 15 minutes! after the fix 3 cycles within 15 minutes! Thanx again!

model: HP Pavillion dv6424 TOSHIBA MK2035GSS sata 200gig

Minardi
May 19th, 2008, 11:32 PM
Thank you very much. I accidentally ran into this topic, and found out that my hd was making 10 cycles in 10 minutes. Thanks to your fix it now stays with 1 cycle in 10 minutes. And after 15 minutes still not moved a cycle. And temp is ~36 degrees. Thanks for this solution, else i needed to buy a new harddisk soon.

My Laptop is a: Acer Aspire 7520-5A2G12Mi -> Definitely Affected!
The Harddisk : Hitachi HTS541612J9SA00 -> Definitely Affected!

Before: 10 cycle | 10 mins -> After fix: 1 cycle | 30 mins

Thats quite a difference :)

packman1234
May 19th, 2008, 11:48 PM
my laptop has the problem also 33 cycles in 15 mins
HP DV6810us

Will this fix affect my system if I dual boot into Vista? I basically use it for Ubuntu only and I unplug it at night and turn it off.
Thanks Nicedude for all the valuable info as usual..Everything I do with Linux works great after using your fixes and advice..I am a newbe.

Nicedude helps us out a great deal and I appreciate him!!!

linuxuser187
May 20th, 2008, 01:20 AM
Also i fixed my friend's pc, cycles before was 48/hour, now 28/hour

it is a:

HP Pavillion dv9804 (Western Digital) WDC WD3200BEVT-60ZCT0 320gig sata

Recruit0
May 20th, 2008, 01:28 AM
My Gateway M-6846 laptop is affected:

WDC WD3200BEVT-22ZCT0: 36 - 38* C

6:55 - 7314
7:15 - 7347

I followed the instructions and the fix did not work.

8:00 - 7402
8:15 - 7425

I'm still getting around 100 per hour. I looked into it and it seems that this is normal for a laptop. I choose to leave it as it is to protect the HD from bumps and reduce heating (although I noticed like maybe a 1* difference in heat, and Ubuntu doesn't seem to support APM or ACPI well for Gateway configurations, attempting to implement a fix isn't that appealing)

Lapino
May 20th, 2008, 01:40 AM
You can change the status of the Dell Inspiron 6400 from "probably affected" to "confirmed affected".

I had an enormous 62 cycles in 15 minutes, but got it fixed.
Thanks a lot!

darco
May 20th, 2008, 02:34 AM
I confirm I have this bug on my Toshiba Satellite A105-S4254.
After running the Load cycle app, I was getting 41 cycles/15mins.

Downloaded fix and now after 15mins, no cycle count increase.

Downloaded and ran hddtmp and I show 33 C

thanks NiceDude!

good luck

darco

hmmm....6 hours running idle my cycle count still hasnt changed?...Is this good or bad? I am now trying to rev up the hard drive (posting this from my laptop,downloading a torrent).
Please advise

darco

KOTAPAKA
May 20th, 2008, 03:17 AM
hmmm....6 hours running idle my cycle count still hasnt changed?...Is this good or bad? I am now trying to rev up the hard drive (posting this from my laptop,downloading a torrent).
Please advise

darco

I have my own way to do this which I think is much easier and equally effective. My counts increase very rarely - 1-2 a day. My HDD is running very smoothly after I made some changes to my system and how it handles the HDD. you can hear no sound at all from the HDD. My apm setting is 255 but temperature stays very low - I'd say lower than in Windows and lower than before I'd addressed this issue.
I might write a how to but I have no time now. I am not saying my way is better than the way presented here. The 2 ways are completely different which is a good thing as some people prefer one way and others rather do it the second or third way. But your count should not increase much especially if you leave your computer idle or just playing music or something like that. I can't be sure but it seems that mine never increases because only thing I do on my laptop is skype, music, internet etc.

werries
May 20th, 2008, 04:40 AM
well thats fun. i thought, oh, this looks scary. and then. OH. OH.....OH NO. My laptop is on that list! I have the dell xps m1210. looks like i'll have to test this soon.

nicedude
May 20th, 2008, 05:55 AM
I am glad to see that I have helped a bunch of people detect and correct this problem, That said I feel like super Mario though as we both work for Stars :KS

To anyone asking about a 0 increase in your cycles count this just means that your system is not parking the hard disk at all which is OK in and of itself assuming you do the temperature test and find that your temp is within the acceptable range for your equipment. The only negative I see with this is a reduction in battery life due to not spinning down when on battery but if that saves your disk from being killed I think its worth it and if I were you I would use the "powertop" utility to try and see what you can do to increase battery life instead. Also try disabling any running services/devices that you don't use ( Blue tooth springs to mind if you don't use it ).

Some have stated that without disk parks being done you are at a risk of damage to your drive if dropped to the floor while my fix ( Or similar ones ) is in place while that could well be true I would add that if you drop your laptop to the ground while it is shut off you could easily destroy it in the process of doing that. Also if your drive is spinning at any given time due to the system using it when you drop it without a fix in place for your system it would be at risk for this as well. So I think that this risk is best mitigated by being careful with your equipment and not dropping it whether running or not & I feel that this bug is more dangerous to someones equipment than the spinning while dropping scenario ( Thanks for pointing it out though as more relevant info is never a bad thing )

I would like some info from advanced users if you do any testing on issues my and other fixes affect. Such as the possible power drain change while on battery that you experience with a fix in place. Also if anyone with advanced knowledge of how this fix could really affect the danger chance of damage to disk during hard bumps as I am assuming currently that if my drive was parking and unparking all the time the danger would be present as well. Of particular interest would be if anyone knows if bumps to the system in the middle of the actual park/unpark operation are even more dangerous than during full spin. If anyone is a hardware specialist with knowledge of this than please PM me with your thoughts on that so I can add them to this thread.

I see 150+ people have downloaded my fix so I am very happy to have possibly saved 150+ hard disks from an early grave :-)

Thanks to all that contribute to helping others. If you have friends using Ubuntu or Linux on a laptop please make them aware of this issue and point them to this or maybe one of the links I list in post #1 so that they can be aware of the issue and protect themselves as well.

Good luck to all of you in all your en devours.

NiceDude

Super Jamie
May 20th, 2008, 09:43 AM
Thanks for this, have starred the original post.

I can confirm the problem is present on my Dell Vostro 1200 with Fujitsu MHY2250BH hard drive (not the standard drive), was getting ~50 cycles per hour.

Cantodea
May 20th, 2008, 10:26 AM
Hi,

My new laptop (1 week old) is also affected, but now thanks to you my laptop is ok. now. Before my harddisk makes every 30 to 50 seconds a strange clicking sound, but now it is gone. However when I start Vista (it's dualboot, Ubuntu 8.04 and Vista on Hdb and Hda) which I don't use very often, the clicking is back. So it seems Vista has this problem to and it is not a Linux thing only. Can I solve the problem in Vista to?

The laptop is a brand new Dell Inspiron 1720, with two HD 500 gigs.The HD vendor seems to be Seagate (ST9250827AS)

Greetings,

J. Beeckaert
Belgium

marcon00
May 20th, 2008, 11:58 AM
Qosmio F40 by default it has Hitachi HTS542512K9SA00 i added to it SAMSUNG HM120JI as it has two HD slots.

Both of them had this problem, thoug Hitachi was getting something close to 1 cycle very 15 min, while SAMSUNG something around 10 !
Hitachi Fixed with -200 settings
meanwhile Samsung -240

Cheers.

nicedude
May 20th, 2008, 12:07 PM
I am most interested in your Vista clicking as you will be the first to confirm for me that they have documented first hand this behavior in Vista. I have a few questions and a request for you cantodea first please answer for me via PM these questions.

1. Do you use MS update to keep your Vista updated or are you using a Pirated copy of vista? ( I could care less about legality and only about figuring it out for you )

2. Has it always clicked since you got it in Vista or did that behavior start at some point?

3. If the behavior started at a certain point can you remember any software you might have installed around that same time ?

4. Did you ever try to use any power management software or change any settings in your BIOS etc ?

Now what I need you to do for me so I can see what is going on with your Vista is please in Ubuntu run the cycle count command to test your number of cycles and write it down. Then use Vista for a set amount of time but best if maybe and hour or more and best if you do something some of the time like web browse or play music etc , then leave Vista and go back to Ubuntu and run the test again and determine the change in the cycle count ( i.e the increase in the number since before using Vista ) Then please PM me by clicking on my name and selecting "send private message" and let me know both the INCREASE and the time period you spent in Vista and I will let you know via PM what this means for your hardware and if it is way too high I will attempt to find a Vista fix for you as well.

Thanks in advance for doing this.

While I really don't want to be giving Vista support here I will try to help anybody here so I will help in this matter for you as well but via PM's :-)

nicedude

KOTAPAKA
May 20th, 2008, 01:16 PM
Funny - I got some clicking as Cantodea has reported. When it happened I was running Debian Etch. My laptop usually makes those clicking sounds but say 1 every 8 hours. I rebooted into Etch one day and it started making them every 30 - 60 seconds or so. Nex day it was gone. In 6 months of running Etch I got this happen twice. Very strange.

GeniusJ
May 20th, 2008, 01:17 PM
I can confirm that this problem also does occour on a new model HP NX7400
with an TOSHIBA MK8034GSX drive.

No clicking sound but a rapid increse in Load_Cycle_Counts.
Once I used the fix the problem was solved, thx!!!

These devices can be added to the startpost.

Rokuta
May 20th, 2008, 03:41 PM
And another confirmation of the problem here. In my case it was on a Zepto Znote 6615WD using a Hitachi HTS721080G9SA00 hard drive. Cycle count increased extremely fast (didn't even have to wait 15 minutes to see the increase was far too high). After applying the fix posted on the first page the cycle count has stopped increasing over the past hour.

Thank you for the wonderful post. I'm glad I'll be able to keep this drive a little longer.

iamvlad
May 20th, 2008, 05:10 PM
Thanks NiceDude!

I had 180 per hour, on a Toshiba Portege 4010. After NiceDude's fix I have 0 for 2 hours :)

Andycas
May 20th, 2008, 05:35 PM
Using HP nx6325 and Samsung HM250JI 250GB. I get 6-8 in 15 min. Fix didnt seem to help either. Guess ill have to use windows till theres a proper fix for it :|

rafafredd
May 20th, 2008, 06:00 PM
I'm using a Brazilian brand Laptop. Megaware (I'm from Brazil), and I get 3 cycles in 15 minutes. Not too bad, but as I see, it could be better. I'm running Heron. Should I try this fix for Heron?

Fenris_rising
May 20th, 2008, 06:32 PM
hi there

saw this post last week and have just got round to having a look at my laptops HDD. my lappy is a budget job (Patriot-Dixons own brand via PC world) sure enough my HDD was cycling up to 24x per 15 mins. implemented your fix and its now 0. thanks very much. incidently the hard drive is an seagate ST94813A there is no bios control of the APM either on this lappy.
i had noticed the HDD did does get bloody hot sometimes so i will monitor that and see what gives, currently its 39C.

EvilRobot
May 20th, 2008, 06:56 PM
Hi guys,

I've just recently started using Ubuntu, well I'm new to Linux in general so I'm not sure if I did this right or not.

I installed and ran smartmontools and the following result was shown the first time:
193 Load_Cycle_Count 0x0032 200 200 000 Old_age Always -248

I then waited 15min. and ran smartmontools again and got this:
193 Load_Cycle_Count 0x0032 200 200 000 Old_age Always -250

Is this good, bad or just plain wrong I know in the first post the instructions say a difference of 2-4 in the last number is ok but my result doesn't look like the example in the first post. Any help would be much appreciated.

nicedude
May 20th, 2008, 07:07 PM
To anyone who thinks like Andycas above that they can avoid this problem by using Windows I am sorry to have to tell you that this problem will affect you in Windows as well.

I have been asking for Vista figures for awhile and a nice user here today did tests and reported the figures to me. He found that he was getting 70 cycles per hour in Vista so don't give up on Linux due to this bug as running back to MS is no solution ( when is it? "uhh never" ding-ding-ding that is correct :-) ).

And if anyone finds that my fix doesn't work for you PLEASE PM ME as Andy is the only one so far to say so and I am trying to get his fixed now so anyone else that suffers from this and yet my fix doesn't correct the problem I will work with you to correct it so please just PM me so I for sure see your in need. If my fix doesn't work for your model I will help you find a solution and would appreciate the info so I can provide solutions for all for this problem.

To all I have helped in this matter I am glad to have done so.

nicedude

ASULutzy
May 20th, 2008, 07:28 PM
Yes, it depends on what Windows power management is set to. Different manufacturers all add and tweak the Windows installs that come with their desktops/laptops.

In Windows I believe just going to the power management section in the control panel and saying never turn off hard disk or whatever may solve the problem of excessive cycling in Windows.

Going to do a more thorough test of this tonight though.

nicedude
May 20th, 2008, 08:10 PM
Well I had intended not to start trying to provide Vista solutions as since Microscum has over $40 Billion in cash to go threatening yahoo with hostile takeovers and an illegal government condoned monopoly they should be fixing their own garbage.

But since it is for you guys I will try to give all Vista users something to try to solve this for when they boot into that. Please just look at the pictures below ( Pulled from web not my machine as Vista is not allowed :-) ) and then go to your drive properties and enable the red circled setting in drive properties as this should stop drive spin down and speed your disk access up as well ( MS warns that by so doing you should never let the power be interrupted suddenly without proper shutdown as data loss can occur, but since this advice is for laptops this should not be an issue ). If anyone does this and would like to help me out please test your load cycles before and after a Vista session and then report the results to me as I am actually interested in the result and as I have stated I don't have vista installed on anything and have no plans to ever use it on any of my boxes, so I have no way to test myself.

nicedude

GeniusJ
May 20th, 2008, 08:25 PM
NiceDude I like you and you are very helpful but, This fix seems so easy and I can't believe the ubuntu coders have not implemented this themselves.

There must be a reason why they haven't ??
Makes you wonder if there are any negative effects of the fix!!

ASULutzy
May 20th, 2008, 08:41 PM
NiceDude I like you and you are very helpful but, This fix seems so easy and I can't believe the ubuntu coders have not implemented this themselves.

There must be a reason why they haven't ??
Makes you wonder if there are any negative effects of the fix!!
Ubuntu doesn't trample all over BIOS/hardware defaults. This is mostly a good thing.

There are some negative aspects of just straight up disabling APM on a hard drive. It will run hotter. Your battery life will decrease considerably.

But for the most part, if you are getting a very very high number of cycles per hour then it's definitely worth at least toning down the number of cycles per hour to something more reasonable.

By default on a lot of machines the power management is way too aggressive.

nicedude
May 20th, 2008, 09:01 PM
Geniusj heres my response - This is the only fix I am aware of ( Not the script itself but the technique it uses aside from using your BIOS to disable APM if setting exists in yours ). I have absolutely no idea why this still plagues most all Linux distros and windows ( which I can now confirm ). Really from my reading its a hardware problem not an OS one so you might try Emailing your disk manufacturer but I doubt you will hear back.

The only negatives I have found are increased power consumption while running from batteries ( How much I don't know as no one has reported any figures and I always use AC and also spend too much time on this as it is ) , increased chance for damage if you drop your laptop while running on to the floor ( Kinda silly since if for any reason its spinning without fix like if in use same applies and dropping your laptop is really dumb in general as they are cheap plastic and can break ) & finally excessive heat due to constant spinning without parking ( which I have yet to get one report of from anyone but would love to hear from anyone that experiences that and by excessive I mean out of the normal acceptable range, as this will most likely increase your disk temp a few "C" but as long as it is within operating specs your all good ).

So since my drive was going to be potentially dead in less than 1 year I an damn glad I found out about this and thought I would pass the info on.

All that said I absolutely agree with those who ask why this bug isn't listed in a sticky or spoken of anywhere that has activity. I think this should be listed somewhere for all laptop users to see but its not and so here we are. I have spent way too much time modding this thread and have spent several hours reading different forums from USA to France and this is best info I know and best fix I could muster.

Please if any Ubuntu developers actually read this I would love to hear from you as well as to why this isn't a sticky or fixed yet or any other relevant info I could pass on.

To all that have helped me by sending me temp and other reports I would like to thank you as you have helped me allot.

nicedude

LeoSolaris
May 20th, 2008, 09:20 PM
I can confirm this bug on a Dell Inspiron 1520 (probably the same hard drive as the 1720)

Thanks! The fix worked for a problem I did not even know I had till I happened on this thread. Probably saved me a lot of head aches down the road.

Leo

By the way, can I delete the script off my desktop after running it? Dumb question I know, but I would like to check first.

reacocard
May 20th, 2008, 09:31 PM
I can confirm this bug on a Dell Inspiron 1520 (probably the same hard drive as the 1720)

Thanks! The fix worked for a problem I did not even know I had till I happened on this thread. Probably saved me a lot of head aches down the road.

Leo

By the way, can I delete the script off my desktop after running it? Dumb question I know, but I would like to check first.

if you've already run the cp commands, then yes, it is perfectly safe to remove the copy on your desktop.

ASULutzy
May 20th, 2008, 10:51 PM
3 posts back you asked about increased temperatures, I can definitely report a noticeable increase in disk temps, I'm usually around 50C now whereas before while I had the constant cycling my temps were more like 44-46C

I'm comfortable with 50C as a hard disk temperature, but others may not be. I've seen linear regression type plots that show the association with temperatures and hard disk life, and the "steepest" point in the curve was from 50-60C, so pretty much anything above 50C starts to be a bit hotter than I really want to run.

werries
May 21st, 2008, 12:18 AM
SEE EDIT:Confirmed bug in Dell XPS M1210, harddrive Hitachi HTS54161.
Fix works as far as reducing the cycles. It was going up about 2-5 cycles per 15 minutes. Now it hasnt moved for an hour. good stuff. So far the temperature has gone from 36 before fix to 39. we'll see how high it gets. but good work. (also, there is a way to reverse this fix if the developers release an official fix, correct? cause heating up worries me.)

EDIT: um. looking at my other temperature control that reads from acpi temperature.
its now at 71 degrees instead of about 60.
EDIT2: Reversed...2-4 cycles per 15 minutes is the lowest range, ends up being almost negligible. Bug NOT confirmed for my laptop and drive in my opinion. but im still a bit new to linux.

narf y akim
May 21st, 2008, 03:43 AM
Yes, I think this must be sticky. I had this problem in my HP laptop and I could solve it (at least that's I think) in january. But before that I had been a long time asking me if that annoing tic.tic.tic, in my laptot produced every few seconds was normal. Finally I found this thread with a so called the ugly fix:

http://ubuntuforums.org/showpost.php?p=3675960&postcount=26

qbashi
May 21st, 2008, 06:50 AM
Hi

My laptop is a HP dv 9000 series to be exact a dv9058ea .
I have the same problems with my hard disk the expected life off it is also 8 months .
But i'am trying openSuse 11.0 beta 3 and there ,there are no problems with it so for me the best way is waiting until the official release off openSuse 11.0.

dirtblack
May 21st, 2008, 08:14 AM
Read your post about the laptops dying early. Before running your script I got around 140 cycles an hour. After I didn't get any. Thanks for your informative post. Oh you can add a Toshiba A135 S2386 to your list.

laoshi
May 21st, 2008, 09:26 AM
I have a Znote 3215W which was affected to some degree.
The test before using the fix had an increase of 10 cycles in 15 minutes.
After fixing: an increase of 0 cycles in 15 minutes!
Thanks for locating and fixing this problem

ditdot
May 21st, 2008, 11:06 AM
nicedude,
i can confirm this bug as well in my acer travelmate 4603 WLCi with Seagate ST9808210A HDD.
After i put your script my hdd now 0 cycle for 1 hour =)

narf y akim
May 21st, 2008, 12:43 PM
Yes, I think this must be sticky. I had this problem in my HP laptop and I could solve it (at least that's I think) in january. But before that I had been a long time asking me if that annoing tic.tic.tic, in my laptot produced every few seconds was normal. Finally I found this thread with a so called the ugly fix:

http://ubuntuforums.org/showpost.php?p=3675960&postcount=26

Hello, yes, I forgot to mention my laptop model. Is an HP 530, and it has been already mentioned. I've bought this one because it was the only one that hadn't had windows pre-installed. I'm really into Ubuntu, but this problem have got me very worried.
Thanks Nicedude for your work in the forum!

tlcrepairs
May 21st, 2008, 12:44 PM
Thanks to nicedude my new lenovo y510 with the listed drive...
/sda: Hitachi HTS542525K9SA00:
Working very smooth after the fix.

Load_Cycle_Count has come to a crawl and runs very cool,
30C and 40C at the most during heavy use.

Before fix: 160 load cycle / hour average :(
After fix: 2 load cycle / hour at most :guitar:

in windoze the load count is still at a high rate but i wont worry
because i rarely use that other os:)

halfmanhalfbug
May 21st, 2008, 01:29 PM
For those with Hitachi drives and brave enough to edit their BIOS settings, Hitachi has a nice tool ("Feature Tool") for changing the default power settings on your HD directly. See:
http://www.hitachigst.com/hdd/support/download.htm?linkto=QL
Download the .iso (and the instructions!), burn it to a CD, and use the CD to boot your computer. The ftool will scan for your drive and you can then go to "Features" and click on "Change Advanced Power Mode" (or press Alt-P). You can change the default using the mouse or the arrow keys. I'm trying it out at "200", we'll see if it melts down.

strick242
May 21st, 2008, 03:28 PM
I can confirm this is a problem on the new Hitachi 250GB serial laptop drive available from Best Buy. This fix also seems to have solved it. My laptop is a Compaq v5207NR. Drive information is as follows:
Hitachi-Travelstar 250GB Serial ATA for laptop
Model# 0A55566

Wikzo
May 21st, 2008, 06:11 PM
My Zepto Znote 3215W appears to have no problems.

I get the same output after 15, 25, 30 and 60 minutes.


wikzo@iBuntu:~$ sudo smartctl -a /dev/sda | grep Load_Cycle_Count
193 Load_Cycle_Count 0x0032 100 100 000 Old_age Always - 2374


And my HDD temperature is always 35C.


wikzo@iBuntu:~$ sudo hddtemp /dev/sda
/dev/sda: TOSHIBA MK8046GSX: 35C

Thank you for the guide! :)

nicedude
May 21st, 2008, 08:57 PM
Thanks to all that are reporting their stats like I asked. But if more could also post their temperature reports as well with their results. Even better would be a pre-fix and post-fix temperature figure while under the same load this would be the most valuable information to me. I am trying to ensure that I am now & can continue through increased knowledge provide the best advice and fix for this problem for all in this community.

Thanks again to all who help in this

napalm brain
May 21st, 2008, 10:24 PM
I have a Sony VAIO VGN-FZ190.
Cycles per hour at 193.

I am kind of flipping out.
The script didn't work for me.

Fenris_rising
May 21st, 2008, 10:49 PM
hi further to my use of your tut my temp goes to 44C tops. sorry no idea what it was before but it does seem cooler. and the cycle count is still stable.

EDIT; been runnning my lappy tonight and as regards temps its deffo running cooler as the fan no longer constantly revs up and down under light duties and im not sweating like a pig because of it either any more.

egruber
May 21st, 2008, 11:41 PM
For those with Hitachi drives and brave enough to edit their BIOS settings, Hitachi has a nice tool ("Feature Tool") for changing the default power settings on your HD directly. See:
http://www.hitachigst.com/hdd/support/download.htm?linkto=QL
Download the .iso (and the instructions!), burn it to a CD, and use the CD to boot your computer. The ftool will scan for your drive and you can then go to "Features" and click on "Change Advanced Power Mode" (or press Alt-P). You can change the default using the mouse or the arrow keys. I'm trying it out at "200", we'll see if it melts down.

I have a Thinkpad T40 with a Hitachi drive and I kept hearing this rapid clicking. I didn't know what it was. Was it caused by this problem? Anyway, I used this utility to reset my APM to 200. I hope that works!

lucipacurar
May 21st, 2008, 11:45 PM
Hey guys

I have and HP 6715s with Hitachi HTS541612J9SA00 and I noticed the "Load/Unload Cycles" problem in Ubuntu and also in Windows XP. My laptop is 3 months old and already has 98286 load/unload cycles :( I'm using my notebook at least 8h a day, most of the time in Windows!


I will add temperatures later.

Edit: Other problems with my notebook here: http://forum.notebookreview.com/showthread.php?t=85477&page=53
:(((

muadnu
May 21st, 2008, 11:59 PM
I had close to 180 cycles/hour on a Dell Vostro 1400...

Just applied the fix, I hope it solves it!

halfmanhalfbug
May 22nd, 2008, 01:27 AM
Hi Nice Dude
Here are my temperature results after running at APM=200:


broose@crash:~$ sudo smartctl -a /dev/sda | grep Temperature
194 Temperature_Celsius 0x0002 122 122 000 Old_age Always - 45 (Lifetime Min/Max 19/52)

Notebook: HP L2000
HD: Device Model: Hitachi HTS541616J9AT00
I have seen it run as hot as 50, but it was about that hot before the fix. Since the linux kernel is logging to the HD several times a minute, the drive never stays spun down for longer than a few seconds even when aggressive power savings are set. I think (and I'm NOT an expert, maybe someone can correct me) that there shouldn't be a big difference in pre-fix and post-fix temperatures.
BTW you can get temperature measurements, including lifetime minimum and maximum, from smartctl. see code above.

nicedude
May 22nd, 2008, 02:17 AM
Thanks for the info halfmanhalfbug I just had someone else tell me about the temp check built into smartctl and will be updating my guide later tonite with the new info. And I think you are correct that the fix should not increase your disk temp that much but I just don't want any chance that I could advise someone to do something that hurts their system and thats why I am preaching the temp check so much. Thanks for all the info and reports as it really helps me stay on top of this problem.

sdennie
May 22nd, 2008, 04:33 AM
I don't think the installation advice in the first post will actually work in Hardy. In Hardy, the resume.d, suspend.d and start.d directories under /etc/acpi no longer work. You'll have to install stuff in /usr/lib/pm-utils/sleep.d or /usr/lib/pm-utils/power.d. I haven't looked at the scripts but, if they are doing something like if on_ac_power, they should work just fine in those two places.

Note:
/etc/acpi/ac.d and /etc/acpi/battery.d still work in Hardy (This is the one change in Hardy that I find really infuriating).

Edit:
I just tried some things on clean Hardy VM's (one updated from Gutsy and one freshly installed). It turns out that /etc/acpi stuff works just fine so, presumably only beta versions of Hardy (even if they are fully updated) are having problems with stuff not working from /etc/acpi. Still, it's probably useful to put the /usr/lib/pm-utils paths in the first post instructions because a lot of people (like me) were probably not patient enough to wait for a final release and won't have most of the directories in /etc/acpi working.

Edit Again:
Actually, I just tested again and I don't think /etc/acpi/suspend.d and /etc/acpi/resume.d are working in Hardy so, the /usr/lib/pm-utils stuff is likely needed.

beyboo
May 22nd, 2008, 05:46 AM
Nicedude,

This thread is getting popular and the inventory getting populated, consider putting a "sorted" inventory list after your main post so that people new to this get to read the problem and can see the increasing inventory below. Just an aesthetic suggestion.

This one on a lighter note to all those tensed about premature deaths here

Several links to offsite information about this bug as the one I got from a member and had posted before was dead.

dead !!! WHO ??? As I read it - the member who probably ran out of his load cylce count ;)

consider punctuation :lolflag:

plamen_todoroff
May 22nd, 2008, 06:29 AM
I found your very recent thread, nicedude, and it seems to be adressing the same issue as http://ubuntuforums.org/showthread.php?t=591503 (which, i don't understand why, is in the archives).

There are some differences between the current thread and the old one, and I don't know which one to follow. The part that bothers me most is the locations to where to copy the script. In the old thread ubuntu_demon says that the script is to be copied in:

$sudo install 99-hdd-ugly-fix.sh /etc/acpi/resume.d/
$sudo install 99-hdd-ugly-fix.sh /etc/acpi/start.d/
$sudo install 99-hdd-ugly-fix.sh /etc/acpi/ac.d/
$sudo install 99-hdd-ugly-fix.sh /etc/acpi/battery.d/

but in here you say it is to be copied here:

sudo cp ~/Desktop/99-sda-load-count-fix.sh /etc/acpi
sudo cp ~/Desktop/99-sda-load-count-fix.sh /etc/acpi/start.d
sudo cp ~/Desktop/99-sda-load-count-fix.sh /etc/acpi/suspend.d
sudo cp ~/Desktop/99-sda-load-count-fix.sh /etc/acpi/resume.d

(I've highlighted the differences)

???

And it gets even more complicated as I stumbled upon another post by the very same ubuntu_demon: https://bugs.launchpad.net/ubuntu/+s...5/comments/232 , in which the locations vary again.

Can you explain about all those locations, as you obviously know more than me. I was about to combine all these locations (yours and ubuntu_demon's) in one bigger command to paste the script in all those places, but I want to know whether it is safe to do so.

And one more question. Your thread says that the fix works in hardy. Why then ubuntu_demon suggest some suse-fix for hardy (http://en.opensuse.org/Disk_Power_Management)? Shouldn't the original ugly fix by ubuntu_demon and your script be used in hardy? They both rely on the same method. The suse-fix seems to apply a different approach.

Please, if someone can clear things up, because my load cycle count issue is closing to 20000 and it's month old laptop, i'd like to stop cycling before it hits 20000.

sdennie
May 22nd, 2008, 06:46 AM
plamen_todoroff: There is no reason to put a script of this nature in /etc/acpi or /etc/acpi/suspend.d. The former will never get run and the latter is setting something that doesn't matter. In Hardy I believe that putting stuff in, /etc/acpi/resume.d also won't get run and you'll instead need to put in a place that pm-utils sees it (I use /usr/lib/pm-utils/sleep.d but, it's possible that /etc/pm/sleep.d will work as well). The /etc/acpi/ac.d and /etc/acpi/battery.d directories are only useful if you want to toggle the hdparm commands depending on whether the laptop is plugged in or not.

nicedude
May 22nd, 2008, 07:49 AM
To vor's points - You are incorrect and my technique works just fine in hardy as I am using hardy right now. I am not 100% that it works with the Hardy beta as I waited for the Release to get a copy, but I see no reason why it would not, but if anyone who is uber guru knows this doesn't work for hardy beta please tell me.

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

To beyboo - I didn't mean that the member died or his drive but I meant that it was brought to my attention that the link was dead ie no longer functioning. So I changed that old link and replaced with 3 working ones instead. All that said I do have reports from 2 members that have reported to me that they have had drives mysteriously die in about a year of 24/7 ( Laptop as a desktop replacement solution ) and believe it is due directly to this. One of the dead drive reports came form a guy who has an extremely high technical grasp of PC's and I think he knows it did for sure. And who knows how many ordinary users would have no idea of this, had drives die and just sent them to be repaired. I do tech work allot on PC's ( Not laptops usually as they have too many screws :-) ) and usually do not do much post-op on why someones stuff broke I just say "this is broke you need a new one it will be this many dollars" and move on so a laptop tech may very well not figure out this killed a drive either ( And if your drive manufacturer figures it out they may keep it too themselves ).

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

To plamen_todoroff I am now aware of the old archived thread ran last year by Ubuntu Demon, when I started this thread I had no idea that there was an earlier one and since that thread has had no activity for quite a while I guess no one else does either. I have sent a PM to Ubuntu Demon saying sorry for starting a thread he already had been modding but I doubt he is going to be mad since both the time since activity on his and since my only motive is to help people here.

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

All that stuff answered I am glad this thread is popular so to speak as I want every single Ubuntu user to read this information so nobody has this bug destroy their disk without their knowledge.

In regards to my fixes effectiveness I can report that as it stands right now 3 people have not had the fix work for them that have reported back to me. first one had high drive temps and I discovered he only had 10-12 cycles per hour and didn't need to use it in the first place - second one had a drive that the 200 hdparm value didn't fix and I gave him different values to try till we fixed it - and the last guy I am helping 1 on 1 right now but after my last value suggestion he has not answered me back yet but I will get his fixed I assure you. To anyone who has any ideas about better advice I can give who knows what they are talking about I am more than happy to listen and implement relevant intelligent changes where needed, and I appreciate the help as well.

Lastly I thought I would share an experience I got today in relation to this thread. I have sent maybe 70-80 PM's randomly to people I saw saying they used a laptop with Ubuntu, in order to ensure some people who obviously use laptops would see this and thereby learn about a possible serious problem. Some of you might have gotten such a PM from me out of the blue and you will remember it said "I am sorry if you already know about this I am sorry for bothering you etc" and was a simple paragraph and a link to this thread. Well today I actually got a response from a person that stated I am a SPAMMER for sending that warning to this individual in question and it wasn't worded nice at all, in fact quite to the contrary.

So I think at this point I am gonna ask everyone I help here to do me ( really the Ubuntu community ) a favor and make sure you try and warn some people and point them to this thread as I would like the thread to go super nova so all can see the problem. I get nothing out of this really except the warm feeling that I helped my fellow humans with a potential problem and I am willing, able & here to help. So please post links to this in laptop discussions or warn laptop users you know here so I don't have to blind PM anyone and thereby don't get responses like I received today. I replied of course that he should report me to the staff and then go look up the definition of Ubuntu and do some soul searching.

Thanks to all that help by posting here your results as its the only way this stays high for all to see, and thanks to all that are providing me with good technical ideas and information I appreciate it.

nicedude

plamen_todoroff
May 22nd, 2008, 08:11 AM
You didn't answer my question where to copy the script and why the given locations differ in different threads? Is it safe to paste the script in every location mentioned in both threads?

sdennie
May 22nd, 2008, 08:14 AM
nicedude: I completely appreciate what you are trying to do but, I really don't think it works over suspend/resume. If you are on a laptop, create a script called 01-touch.sh with the following:



#!/bin/bash

touch /tmp/it-ran


Put it in /etc/acpi/resume.d with:



sudo install 01-touch.sh /etc/acpi/resume.d


Suspend and resume your machine. Then check /tmp to see if a file "it-ran" exists:



ls /tmp


Now move the script so that pm-utils sees it:



sudo mv /etc/acpi/resume.d/01-touch.sh /usr/lib/pm-utils/sleep.d


Suspend, resume and again check /tmp.

nicedude
May 22nd, 2008, 08:53 AM
plamen_todoroff - you could copy my script to every folder on your machine and it would do nothing negative to it.

vor - out of 200+ people using my fix succesfully your the first to suggest this is occuring. I will try to look into what you are saying but I am pretty busy and have been modding this thread as a service for others and it has been more like a no-pay part time job the last week to do this.

I will have to ask someone else to test this though as I like many other laptop users have problems with Hardy and Hibernation and rather than worry with spending time figuring out why at this junction (I will just wait for more people to work on why that is) I instead don't use hibenation and instead just have my system set to shut down at a certian battery level just in case I would forget and leave it on somewhere while on batt. For the most part ( 95% ) I use mine as a desktop PC anyway and as such Hibernation mode really has little value to me. So if someone actually does the work and finds that what you say will occur does in fact occur and can suggest a modification to the script or additional locations it needs to be copied to I will reconfigure my instructions for that. If I get you correctly you are stating that while my fix works it won't work once the PC is resumed from a Hibernation session by the way are you not?

PS I got the locations for where to copy the script to from a offsite thread about this problem concerning Ubuntu and have confirmed that they are the same locations listed at several other offsite discusions on this problem and up to now you are the first person to say they are wrong.

If you would do the testing and confirm that what you say is correct or not I would certainly listen to your findings with an open mind and if they prove true I will try to incorporate them in the guide.

Thanks for any work you do to contribute, please report your findings via PM to me so we can further this discussion in that way.

plamen_todoroff
May 22nd, 2008, 09:02 AM
I'm NOT saying in any way that your locations are wrong, just couple of them are different from what I've read. I really appreciate your work and efforts. I just wanted to know something more about the possible locations to copy the script and why they are different in different posts. No bad feelings. Thank you for the script.

sdennie
May 22nd, 2008, 09:24 AM
nicedude: Again I respect what you are trying to do but, it's not quite correct. A few things:

1) Putting a script in /etc/acpi will do nothing.

2) Putting a script to modify hdparm settings in /etc/acpi/suspend.d will do nothing (It will set the hdparm just before it goes to sleep and immediately lose those setting upon waking up).

3) Setting the APM on a disk is actually useful on a laptop. I haven't looked at your scripts but I assume they are blindly setting hdparm -B to 254 or 255. Using something similar to the last link in my signature will disable the annoying click (and possible disk failure) on AC power, while still allowing you to save power while on battery (statistically, I think it's impossible to kill a laptop via Load_Cycle_Count if it's only enabled while on battery power (unless you have a lot of extra batteries))

I like what you are doing (though, I admit that I don't approve of the gloom and doom title of the thread) but, please understand all the details of what you are doing.

nicedude
May 22nd, 2008, 10:24 AM
As I said previously I got these locations from research on this in other off site threads concerning Ubuntu and if putting it somewhere does nothing it also hurts nothing, And as I found these locations on other multiple threads by people more knowledgeable than me I won't be changing anything due to people who post what they think without first even looking in the script to see whats there as you did in your first post. I will change anything that is incorrect when people who know what they are saying and have tested their hypothesis let me know of their findings. If you don't think my script works or would rather fix yours yourself another way then by all means do so as this is intended to warn people of a potentially bad situation ( Who may have no idea about Ubuntu and be brand new to it ) and thats it. I asked you to test your hypothesis and report findings to me via PM and instead you do nothing but just repost more negative assessments this time even of the title of my thread, why don't you actually do some research on this instead of judging my actions as I really don't care what you think of me. And you can save the I really appreciate you as wrapping negative comments with a veil of niceness is just known as a clever insult where I am from and I just listen to people who actually both know what they are talking about and research their ideas before spouting them out as facts.

So once again for all here this script is not dangerous to copy anywhere on your system so if a location does nothing then it hurts nothing as well and if someone uses battery mode all the time then they are load cycling the drive constantly still and it could affect their drives lifespan. Its up to them as I state all over the place to do the math and see what lifespan is acceptable to them.

To all that offer kind insight and advice I thank you as you guys have helped me allot

sdennie
May 22nd, 2008, 10:56 AM
I was actually being sincere when I said that I respected what you are trying to do. I downloaded your script and it's (not surprisingly) just an hdparm -B command. So, again, your installation steps are not valid. Just because installing something does not produce immediate and detrimental effects doesn't mean it's correct. The spirit of this thread is great but the implementation needs a bit of work (not an insult, I mean that you need to understand exactly how parts of the system are interacting before you can give really good advice as to how those interactions should change).

jacklsw2008
May 22nd, 2008, 11:49 AM
i wonder mine could be considered suffering from the bug or not?

Thu May 22 18:17:22 MYT 2008
193 Load_Cycle_Count 0x0032 199 199 000 Old_age Always - 5825
194 Temperature_Celsius 0x0022 097 092 000 Old_age Always - 50

Thu May 22 18:25:14 MYT 2008
193 Load_Cycle_Count 0x0032 199 199 000 Old_age Always - 5846
194 Temperature_Celsius 0x0022 098 092 000 Old_age Always - 49

Gilson
May 22nd, 2008, 12:30 PM
My results:

193 Load_Cycle_Count 0x0032 099 099 000 Old_age Always - 17201
194 Temperature_Celsius 0x0022 100 100 000 Old_age Always - 49 (Lifetime Min/Max 9/55)


On a Acer Aspire 3690, using Ubuntu Hardy Heron.

rubenlj
May 22nd, 2008, 03:48 PM
My Toshiba Satellite P100-103 laptop with Ubuntu Hardy is affected too. The nicedule's fix is working properly.

derfie
May 22nd, 2008, 05:13 PM
Many clicks on Thinkpad T41 with Hitachi HTS541040G9AT00 too.
Clicks are gone after fix and restart.
BUT after resuming from sleep or hibernation same problem occurs again.
Any idea what coarses this ? All scripts are copied in the right places with execution rights.

Doombringer
May 22nd, 2008, 05:40 PM
I have a Dell Inspiron 1520, I am affected by the bug but the fixes dosen't work for me... it keeps increasing very fast.

But, if I run the command manually, it works. It looks like the script dosen't run on startup or something. I did check "Allow executing as program" but still, it does not work :(

genwei_0
May 22nd, 2008, 06:11 PM
Hi, I used ur script to lower the load cycle count, thank you very much. Here is the information you request in ur thread:
Model of laptop & hard disk: Compaq Evo N1000v(I bought it 6 years ago) & Seagate ST96812A(I replaced the original 20GB drive with this 60 GB one 2 year ago)
1 hour cycle count before fix & after: before fix, the count is 19x4 = 78/hour; after fix, the count is zero(is it normal?)
Temprature before fix & after fix: before fix, the temp is fluctuating between 38-40C; after fix, the temp still stays in between this range.

thanks again for saving my hard drive.:guitar:

egruber
May 22nd, 2008, 06:22 PM
Many clicks on Thinkpad T41 with Hitachi HTS541040G9AT00 too.
Clicks are gone after fix and restart.
BUT after resuming from sleep or hibernation same problem occurs again.
Any idea what coarses this ? All scripts are copied in the right places with execution rights.

Why don't you try the Hitachi utility referenced earlier in the thread. Tha's what I used. It's much easier than executing scripts, and it saves the changes directly to the device.

Randomiser
May 22nd, 2008, 06:36 PM
Thankyou mate this helped alot, my drive was affected by this.
My drive is the WDC WD1200BEVE-0 and was doing around 150 counts in an hour.
The only way i could i could significantly stop this was by setting the -b value to 255 my drive temp is 34c at high loads and 32 at minimal.

Thanx Nicedude.:KS

lswest
May 22nd, 2008, 07:28 PM
HP DV6545EG - according to device number (googled it) the hdd is: Seagate Momentus 5400.3 ST9160821AS 160GB 5400 RPM 8MB Cache SATA 1.5Gb/s Notebook Hard Drive - OEM

load count (averaged from a few 15 minute tests for the 1 hour) ~120 load cycle increase.

after fix:

15 minute tests, increase: 0, 1 hour increase ~1-2

side note: i noticed that without the hdparm at 200 my load count would increase by 1 every time i skipped/started a new song in Exaile, might have led to the excessive amount of increase.

Temp: 47 +/-3 (no change)

derfie
May 22nd, 2008, 08:03 PM
Thanks Egruber for the suggestion but that's not what I want. System is dualboot and that would mean no disk powermanagement under windows too.
When I start or restart the machine the script is working well,so the fix is working, but when I resume the machine it just forgets the settings. Strange isn't it ........

troupa
May 22nd, 2008, 10:34 PM
affected: HP Pavilion dv9617nr

cycles before fix ~162
temp before fix sorry did not note this
cycles after fix None seen in over an hour
temp after fix 43

Thank you :)

EDIT I watched a movie for 30 minutes to keep the hard drive spinning to check temp. after that it went from 43 to 46

ky4623
May 23rd, 2008, 03:46 AM
Dell E1505, Model=TOSHIBA MK8034GSX:

before the fix the count increased by around 20 in 5~6 mins.

after the fix:
Thu May 22 22:15:40 CDT 2008
193 Load_Cycle_Count 0x0032 093 093 000 Old_age Always - 70742
194 Temperature_Celsius 0x0022 100 100 000 Old_age Always - 36 (Lifetime Min/Max 8/56)

Thu May 22 22:30:40 CDT 2008
193 Load_Cycle_Count 0x0032 093 093 000 Old_age Always - 70745
194 Temperature_Celsius 0x0022 100 100 000 Old_age Always - 35 (Lifetime Min/Max 8/56)

rafttrip
May 23rd, 2008, 08:40 AM
Affected model VGN-CR125E

My Terminal output

Sat May 24 05:27:56 WST 2008
190 Temperature_Celsius 0x0022 062 049 045 Old_age Always - 639828006
193 Load_Cycle_Count 0x0032 060 060 000 Old_age Always - 81008
194 Temperature_Celsius 0x0022 038 051 000 Old_age Always - 38 (Lifetime Min/Max 0/16)

Sat May 24 05:47:59 WST 2008
190 Temperature_Celsius 0x0022 064 049 045 Old_age Always - 639828004
193 Load_Cycle_Count 0x0032 060 060 000 Old_age Always - 81064
194 Temperature_Celsius 0x0022 036 051 000 Old_age Always - 36 (Lifetime Min/Max 0/16)


that was befor fix

this is output after fix

Sat May 24 06:19:19 WST 2008
190 Temperature_Celsius 0x0022 064 049 045 Old_age Always - 606273572
193 Load_Cycle_Count 0x0032 060 060 000 Old_age Always - 81145
194 Temperature_Celsius 0x0022 036 051 000 Old_age Always - 36 (Lifetime Min/Max 0/16)

Sat May 24 06:42:16 WST 2008
190 Temperature_Celsius 0x0022 064 049 045 Old_age Always - 606273572
193 Load_Cycle_Count 0x0032 060 060 000 Old_age Always - 81145
194 Temperature_Celsius 0x0022 036 051 000 Old_age Always - 36 (Lifetime Min/Max 0/16)


Thank you for your hard work. I very much appreciate the effort.
Hopefully this will be fixed in an update or some thing like that soon.:)

Fenris_rising
May 23rd, 2008, 10:43 AM
previously posted in here about the successful use of this guide on my laptop.still stable and cooler running. it just makes me wonder if my original hard drive under the pre-installed XP OS suffered from the problem that this guide has cured. it lasted 14 months with minimal use, luckily i had an insurance with PC world on it so it cost me nothing to fix. but if i hadnt i may well have been 400 worse off.

troupa
May 23rd, 2008, 10:50 AM
Fenris:

My laptop, due to issues, is still dual boot. I was using Windows Vista (yuck!) earlier and when I rebooted into Ubuntu I checked the drive and it had shot up 70 cycles while in windows for ~20 min. (well... 68... one for each reboot I guess.) And that is with the power settings in Vista set supposedly to never shut down the hard drive. Ah another lie from M$... their power "management." Heh. I sure won't be using Vista unless I absolutely HAVE to.

So, probably.

Fenris_rising
May 23rd, 2008, 10:59 AM
hmmmmm. do you ever get the feeling that 'built in obsolescence' is a being driven by the manufacturers rather than the limits of the materials used?

"i smell rich pickins my lovely" < fagin like CEO of a major manufacturer

plamen_todoroff
May 23rd, 2008, 11:55 AM
Why I don't get the minimum and lifetime temperature readings with this command: date; sudo smartctl -a /dev/sda | egrep '(Load_Cycle_Count|Temperature)'? I get this only:

Fri May 23 13:54:24 EEST 2008
193 Load_Cycle_Count 0x0032 194 194 000 Old_age Always - 19939
194 Temperature_Celsius 0x0022 098 095 000 Old_age Always - 49

And there is nothing close to it here also: sudo smartctl -a /dev/sda.

What could be the problem?

revelationman
May 23rd, 2008, 12:50 PM
I have ran the test many times this is on a HP NX6325 and the issue for premature hard drive was evident.

I am waiting for a fix since that I have reverted to back to Windows.

I hope there will be a fix and when there is I will be putting in my Hardy Disc right away.

Just a quick note I still run Hardy on my 11 year old Compaq Deskpro without any issue's, however this is a IDE drive not SATA.

:guitar:

lucipacurar
May 23rd, 2008, 05:52 PM
you should check this issue in Windows too.

Ripfox
May 23rd, 2008, 06:52 PM
nicedude: Again I respect what you are trying to do but, it's not quite correct. A few things:

1) Putting a script in /etc/acpi will do nothing.

2) Putting a script to modify hdparm settings in /etc/acpi/suspend.d will do nothing (It will set the hdparm just before it goes to sleep and immediately lose those setting upon waking up).

3) Setting the APM on a disk is actually useful on a laptop. I haven't looked at your scripts but I assume they are blindly setting hdparm -B to 254 or 255. Using something similar to the last link in my signature will disable the annoying click (and possible disk failure) on AC power, while still allowing you to save power while on battery (statistically, I think it's impossible to kill a laptop via Load_Cycle_Count if it's only enabled while on battery power (unless you have a lot of extra batteries))

I like what you are doing (though, I admit that I don't approve of the gloom and doom title of the thread) but, please understand all the details of what you are doing.

This is correct.

nicedude
May 23rd, 2008, 06:52 PM
lucipacurar is absolutely correct as since this is a hardware issue and not a software issue Windows is affected by this as well unless you change your power management settings in Windows which will in effect do the same thing that the Ubuntu fixes do and disable drive head parking. I have last night after doing more research changed my suggested locations a bit for my script fix thanks to vor and another kind person who sent me a PM to inform me he wasn't being protected after resuming from hibernation ( sorry but hibernation was buggy on my laptop with hardy last time I tried and so I just don't use it since I use my laptop like a desktop so I was unaware of this ). I have also added 2 commands to remove my script from the 2 places it is not needed ( But they will not have any negative affect if left in those locations they just wont have any positive effect either hence the commands to remove them if you feel you want to ). And I have also added 2 places for the 2 I removed that should allow the fix to work after a resume from hibernation and my copy commands now reflect those so if you just rerun the copy commands with the script fix for how your drive is addressed (sda or hda) then the new locations will get the script and you should have a working fix even post hibernation. Sorry to everyone for the extra trouble this may cause but I am trying to provide the best solution I can while at the same time trying to make sure it works for everyone all with only one laptop to test anything on and not hibernating myself ( also I am far from being perfect quite the opposite in fact ). I hope you all find this thread helpfull to you and I hope everyone likes Ubuntu as much as I do.

On the table for this thread as time allows are 1. trying to figure out a fix that would still save more battery by allowing spindown but at a safer rate while in battery mode 2. Alphabetically sorting the list of affected systems and drives as soon as I get a chance to. If anyone find relevant info on this topic I dont address please PM me with the details and I will investigate/implement changes as needed.


Thanks to all that help and good luck in all you do,

nicedude

Ripfox
May 23rd, 2008, 06:57 PM
nicedude: I completely appreciate what you are trying to do but, I really don't think it works over suspend/resume. If you are on a laptop, create a script called 01-touch.sh with the following:



#!/bin/bash

touch /tmp/it-ran


Put it in /etc/acpi/resume.d with:



sudo install 01-touch.sh /etc/acpi/resume.d


Suspend and resume your machine. Then check /tmp to see if a file "it-ran" exists:



ls /tmp


Now move the script so that pm-utils sees it:



sudo mv /etc/acpi/resume.d/01-touch.sh /usr/lib/pm-utils/sleep.d


Suspend, resume and again check /tmp.

+1

daberger
May 24th, 2008, 06:01 AM
i report that the compaq presario 2594 US has this porblem


unfortunately this is a hand me down laptop (craptop?) and has been well past its life time with extremely low performance for a long time now. i have almost 670000 loads!

Metaleks
May 24th, 2008, 06:16 AM
Not a problem, but I get two temperature readings. Odd.


190 Temperature_Celsius 0x0022 062 028 045 Old_age Always In_the_past 706871334
193 Load_Cycle_Count 0x0032 096 096 000 Old_age Always - 8678
194 Temperature_Celsius 0x0022 038 072 000 Old_age Always - 38 (Lifetime Min/Max 0/19)

Also, what would an average estimate be of hard drive temperatures? When doing heavy stuff (copying gigs of files) I get about 39-40 degrees. Never more. It's usually pretty consistent at 38. Should I be worried? The clicking has stopped (I used to count 18 a minute), but has it stopped at the cost of a higher temperature?

daberger
May 24th, 2008, 06:35 AM
and a quick question, how do you download the script. i tried to apt-get install but it didnt work

-newbie

sam_delta
May 24th, 2008, 08:29 AM
download them from the first post of this thread, at the end of the post, click on either script you want to download and follow the instructions on the first post on how to install it

sam

hotweiss
May 24th, 2008, 08:58 AM
I can tell you that the Seagate 5400.4 250 GB is affected. I've got it in my Asus M50Sv-A1. I went through 12 cycles in 5 minutes and about 40 cycles in 15 minutes.

I've applied your patch and after 15 minutes I didn't even complete one cycle. My hard drive stopped ticking to! Thank-you very much! This is a rather serious problem. Shouldn't Ubuntu up the patch right away?

PS-does this problem affect Vista to?

unimatrix
May 24th, 2008, 10:03 AM
Model: Compal FL90
HDD: NOT affected

Output:


user@hostname:~$ date; sudo smartctl -a /dev/sda | egrep '(Load_Cycle_Count|Temperature)'
Sat May 24 10:43:42 CEST 2008
193 Load_Cycle_Count 0x0032 159 159 000 Old_age Always - 123602
194 Temperature_Celsius 0x0022 120 097 000 Old_age Always - 27
user@hostname:~$ date; sudo smartctl -a /dev/sda | egrep '(Load_Cycle_Count|Temperature)'
Sat May 24 10:58:45 CEST 2008
193 Load_Cycle_Count 0x0032 159 159 000 Old_age Always - 123602
194 Temperature_Celsius 0x0022 115 097 000 Old_age Always - 32

derfie
May 24th, 2008, 11:12 AM
Hi Nicedude,
Appreciate the work you did and still do.
Have tried your suggestion to copy the scripts in the two new locations (ac.d and battery.d) but it did not solve the problem of resuming. (still same number of clicks after resuming from sleep and hibernation on thinkpad T41 with Hitachi drive)
Have a nice day !

barnex
May 24th, 2008, 01:52 PM
Thanks for saving my drive, you're a very nice dude, nicedude ;) Anyways, my machine had a load cycle of about 1 / minute before the fix, and zero now (I put hdparm -B 245). The temperature has gone up from 37C to 41C. It's a first generation apple macbook with a Toshiba MK603 hard drive. I had already found this kind of fix in a thread specifically about macbooks, but there they forgot to put the script in /etc/acpi/ac.d (only in battery.d). I'll see if I can find that thread and point those poor macbook owners to your thread.

hotweiss
May 24th, 2008, 02:19 PM
Hmmm, there's only one problem. When I suspend my computer and then turn it back on, I go back to the old hdparm setting.

barnex
May 24th, 2008, 02:24 PM
Hmmm, there's only one problem. When I suspend my computer and then turn it back on, I go back to the old hdparm setting.

Confirmed, exactly the same here: after waking from suspend I go back to my usual 1 cycle / minute. Let's see if this gets fixed by going to hibernation and back...

-------------EDIT----

nope, after hibernation it's still on about 1cycle/minute.

hotweiss
May 24th, 2008, 02:37 PM
Confirmed, exactly the same here: after waking from suspend I go back to my usual 1 cycle / minute. Let's see if this gets fixed by going to hibernation and back...

-------------EDIT----

nope, after hibernation it's still on about 1cycle/minute.

I just ran smartmon in Vista, and the problem is non-existent there, leading me to believe that this is not a firmware issue.

rlbarket06
May 24th, 2008, 03:04 PM
Thank you for your great post!

My setup:
Dell Inspiron 6400 Laptop
Aftermarket Toshiba HDD 120GB 7200RPM
Ubuntu 8.10

Results:
BEFORE:
Count/Hour - 150
Temp - 39

AFTER
Count/Hour - 0
Temp - 45

Unfortunately, this brand new HDD already went through 11% of its expected cycles before I found your post! Thanks again.

tmcmack
May 24th, 2008, 03:36 PM
Holy cow! I figured that this was only when using the battery, but my laptop went through 18 cycles in 15 minutes on A/C power!

EDIT: Now the cycle count hasn't changed in 3 hours of light use (web browsing, youtube, some internet radio). Is that a problem?

Before:

~
$ date; sudo smartctl -a /dev/hda | egrep '(Load_Cycle_Count|Temperature)'
Sat May 24 09:31:51 EDT 2008
193 Load_Cycle_Count 0x0012 068 068 000 Old_age Always - 326610
194 Temperature_Celsius 0x0002 144 144 000 Old_age Always - 38 (Lifetime Min/Max 9/57)

~
$ date; sudo smartctl -a /dev/hda | egrep '(Load_Cycle_Count|Temperature)'
Sat May 24 09:45:40 EDT 2008
193 Load_Cycle_Count 0x0012 068 068 000 Old_age Always - 326628
194 Temperature_Celsius 0x0002 144 144 000 Old_age Always - 38 (Lifetime Min/Max 9/57)


Afterwards, no change in 30 minutes:

~
$ date; sudo smartctl -a /dev/hda | egrep '(Load_Cycle_Count|Temperature)'
Sat May 24 10:02:13 EDT 2008
193 Load_Cycle_Count 0x0012 068 068 000 Old_age Always - 326640
194 Temperature_Celsius 0x0002 137 137 000 Old_age Always - 40 (Lifetime Min/Max 9/57)

~
$ date; sudo smartctl -a /dev/hda | egrep '(Load_Cycle_Count|Temperature)'
Sat May 24 10:35:38 EDT 2008
193 Load_Cycle_Count 0x0012 068 068 000 Old_age Always - 326640
194 Temperature_Celsius 0x0002 144 144 000 Old_age Always - 38 (Lifetime Min/Max 9/57)



This is a 4.5 year-old lappy, so I will hope that 325k load cycles is roughly proportional to its lifespan, despite the fact that it doesn't get heavy use and I don't use it on battery power very often. Although, I've had Ubuntu on it for more than a year -- wonder how long this has been going on... Time to run a backup.

Toshiba Satellite A25-S279
Ubuntu Hardy
Not sure of the HDD specs; it's all stock parts with a 60GB drive.

barnex
May 24th, 2008, 03:53 PM
Fixed keeping the settings after suspend/hibernate

I've made a temporary fix so that the correct settings are 'kept' after suspend or hibernate. It's very ugly, but it works. I've simply created a crontab that re-applies the correct settings every 2 minutes.

Here is how I fixed it: Edit the crontab file for root:


sudo crontab -e


An editor will open which allows you to edit the crontab file.
Add the following line:


*/2 * * * * /sbin/hdparm -B 245 /dev/sda


If you want to check that this really works, you can add this line instead, this will log the activity.


*/2 * * * * date>> /home/user/Desktop/log; /sbin/hdparm -B 245 /dev/sda >> /home/user/Desktop/log


Save the file and exit the editor. The power settings should now be applied every 2 minutes, even after suspend or hibernate. To check if it really works, the log file on your Desktop should look like this:


Sat May 24 16:42:01 CEST 2008
/dev/sda:
setting Advanced Power Management level to 0xf5 (245)

Sat May 24 16:44:01 CEST 2008
/dev/sda:
setting Advanced Power Management level to 0xf5 (245)

If all works well, you can disable the logging (put the fist line in the crontab instead of the second line). You can also replace the */2 by e.g. */5 to re-apply the settings only every 5 minutes.

I know this is an ugly solution, but it's keeping my harddrive from dying at this moment (I'm already more than halfway the maximum number of load/unload cycles).

Hope this helps. Good luck all!

DaveTheAve
May 24th, 2008, 05:16 PM
Hey, I've got it working on mine THANKS! Though, I do have an issue, read below:

Model: Toshiba Satellite A135-S4467
OS: Kubuntu 8.04 KDE 4.0.3

Before:
20min Running -> 44 cycles (~132/hr)
Temp: 38C

After:
1hr Running -> 0 cycles
Temp: 49C

Question/Problem:
Is it good that my HDD is running at 49C under idle? Please email me at David _@_ NeoeliteUSA D0T com about this. THANK YOU.

brharri1
May 24th, 2008, 07:00 PM
My Acer Aspire 5570z was affected and the script took care of it. Thank you very much.

sam_delta
May 24th, 2008, 07:22 PM
I just ran smartmon in Vista, and the problem is non-existent there, leading me to believe that this is not a firmware issue.

because vista reads the hdd much much more than ubuntu, it is almost ALWAYS reading from the hard drive, so it never parks

an ubuntu developer explained this some time ago, i dont remember where i read it from, but he explained that its not a ubuntu problem but a firmware/bios problem,

firmware and bios settings can be override by the operating system, and thats what the fix should be about. but Ubuntu by default, just uses the default manufacturer settings (bios/firmware) ,which in fact SHOULD be the best for your hardware (which might not be the case)

sam

bigbrovar
May 24th, 2008, 09:19 PM
Model: Sony Vaio FZ21E
OS: Ubuntu 8.04 Gnome


1 hour cycle count before fix ->40 cycles
Temp: 43C

1 hour cycle count after fix ->0 cycles
Temp: 42C


worked like a charm... thanks for the great work u are doing for the community ..

bapoumba
May 24th, 2008, 09:39 PM
http://ubuntuforums.org/showthread.php?p=5030999

There is a sticky thread in H&L (it's been copied from an older one that was in the archive after the forums software upgrade, here (http://ubuntuforums.org/showthread.php?t=591503))
It can be quite confusing to have several threads running on the same issue. ubuntu-demon has been handling this from the start. Please read the warnings, the various informations, the bug reports, how to revert the applied fixes etc. and contribute to the sticky, thanks.

I'm closing here, sorry.