Post new topic   Reply to topic
View previous topic Printable version Log in to check your private messages View next topic
Author Message
darioOffline
Post subject: Load Cycle Count value increasing very fast  PostPosted: 12.04.2011, 12:09



Joined: 2010-11-27
Posts: 63

Status: Offline
I noticed that the Load_Cycle_Count value is increasing very fast in my Acer Extensa 5220 laptop:
      Code:

Host/Kernel/OS  "pc-luisa" running Linux 2.6.38-2.slh.6-aptosid-686 i686 [ aptosid 2010-03 Ἀπάτη - kde-lite - (201012262151) ]
CPU Info        Intel Celeron 560 @ 1024 KB cache flags( sse3 nx lm ) clocked at [ 2128.207 MHz ]
Videocard       Intel Mobile GM965/GL960 Integrated Graphics Controller (primary)     [  ]
Network cards   Broadcom NetLink BCM5787M Gigabit PCI Express
                PCI device 0x4243:0x4315
Processes 111 | Uptime 1:07 | Memory 256.0/2016.9MB | HDD Hitachi HTS54321 Size 160GB (44%used) | Client Shell | Infobash v3.35


The second value is read a minute after the first:

      Code:

smartctl -a /dev/sda | grep Load_Cycle
193 Load_Cycle_Count        0x0012   087   087   000    Old_age   Always       -       130655

second:
      Code:

smartctl -a /dev/sda | grep Load_Cycle
193 Load_Cycle_Count        0x0012   087   087   000    Old_age   Always       -       130672


Is there any way to definitively stop this?

Update: searching, i found the advice to add:
      Code:

/dev/sda {
    apm = 254
    spindown_time = 0
}

in /etc/hdparm.conf, i tried to do this but the value is still increasing.
Instead, the command
      Code:

hdparm -B 254 /dev/sda

works fine and the value stops increasing, what do i have to do to execute this command at every system start?
 
 View user's profile Send private message  
Reply with quote Back to top
broweOffline
Post subject: RE: Load Cycle Count value increasing very fast  PostPosted: 12.04.2011, 16:15



Joined: 2010-09-12
Posts: 152
Location: Canada
Status: Offline
As root try this:

      Code:

crontab -e
@reboot hdparm -B 254 /dev/sda


As a general note, its a good idea for everyone to keep an eye on this value whenever they get a new HD. I have a WD GreenDrive that I'm using as a primary drive (yeah I know its not recommended, but it's been working 24/7 for 2yrs now) and some models have a similar problem with excessive head parking / load cycle counts.
 
 View user's profile Send private message  
Reply with quote Back to top
darioOffline
Post subject:   PostPosted: 12.04.2011, 16:37



Joined: 2010-11-27
Posts: 63

Status: Offline
I have a WD Caviar green as main drive on another pc, the load cycle count value is high (30905 in less than two months).
I didn't find any way to stop this, the hdparm -B * commands aren't supported:
      Code:

# hdparm -B 254 /dev/sda

/dev/sda:
 setting Advanced Power Management level to 0xfe (254)
 HDIO_DRIVE_CMD failed: Input/output error
 APM_level      = not supported


I have this drive because of a misunderstanding with the shop, i think i'll buy an SSD as soon as possible.

p.s. what does "@reboot" mean? I never used the crontab command before
 
 View user's profile Send private message  
Reply with quote Back to top
ShakaZOffline
Post subject:   PostPosted: 12.04.2011, 19:34



Joined: 2011-02-16
Posts: 67

Status: Offline
I guess my hard drive must be really bad than :p.
      Code:
# smartctl -a /dev/sda | grep Load_Cycle
225 Load_Cycle_Count        0x0032   001   001   000    Old_age   Always       -       7035168

This laptop is barely over 1 year old.
 
 View user's profile Send private message  
Reply with quote Back to top
darioOffline
Post subject:   PostPosted: 12.04.2011, 22:12



Joined: 2010-11-27
Posts: 63

Status: Offline
the value is very high, which is the laptop model?

For the hdparm command, i also found this script in Arch Linux wiki:
      Code:

#!/bin/sh
 
if [ -n "$1" ] && ([ "$1" = "resume" ] || [ "$1" = "thaw" ]); then
   hdparm -B 254 /dev/your-hard-drive > /dev/null
fi


this script is needed to execute the command after resume from suspension. Will this script work in debian too if put in /etc/pm/sleep.d/?
 
 View user's profile Send private message  
Reply with quote Back to top
ShakaZOffline
Post subject:   PostPosted: 13.04.2011, 10:09



Joined: 2011-02-16
Posts: 67

Status: Offline
It's an Acer Aspire 5738ZG. I received it in January last year. The laptop is actually a bit older than that, but the guy who gave it to me swapped it's 500 GB hard drive for the 300GB that came in his new one. He bought the new one because this one has a glassy screen that is very difficult to read in the sun, and he was touring the world on his sail boat.

The hard drive is a Samsung SpinPoint M7 model HM320II according to smartctl.
 
 View user's profile Send private message  
Reply with quote Back to top
darioOffline
Post subject:   PostPosted: 13.04.2011, 10:42



Joined: 2010-11-27
Posts: 63

Status: Offline
Did you try to use hdparm -B (value > 128) to stop the spindown?

for example:

      Code:

hdparm -B 128 /dev/sda #replace with your device name


Values over 128 will theoretically stop the spindown, actually some hard drive need higher values, like 200, 254, 255 (which means no powersave). I experienced this with both the acer extensa and an hp compaq nx7400.
 
 View user's profile Send private message  
Reply with quote Back to top
ShakaZOffline
Post subject:   PostPosted: 13.04.2011, 12:05



Joined: 2011-02-16
Posts: 67

Status: Offline
No I haven't. According to this http://lists.debian.org/debian-laptop/2 ... 00073.html a high Load_Cycle_Count is actually a good thing as it protects the drive from data loss.
The more relevant Start_Stop_Count value that tracks spin-up/spin-down is astonishingly quite low at 623 as laptop drives are supposed to handle half a million of those. Maybe that is because hdparm reports advanced power management as disabled on this drive, which I guess means the drive only spins up and down when powering the laptop or shutting it down.
 
 View user's profile Send private message  
Reply with quote Back to top
darioOffline
Post subject:   PostPosted: 13.04.2011, 12:46



Joined: 2010-11-27
Posts: 63

Status: Offline
Mm.., i'm not completely convinced of that, it is the first positive thing i read about this topic, i'll search more.
 
 View user's profile Send private message  
Reply with quote Back to top
diblOffline
Post subject:   PostPosted: 13.04.2011, 13:00



Joined: 2010-09-12
Posts: 302
Location: Dayton, Ohio, USA
Status: Offline
Here's the best information I know of, on the topic: https://ata.wiki.kernel.org/index.php/K ... nder_Linux
 
 View user's profile Send private message Visit poster's website  
Reply with quote Back to top
ShakaZOffline
Post subject:   PostPosted: 13.04.2011, 13:32



Joined: 2011-02-16
Posts: 67

Status: Offline
Hmmmm... then either I have an uber hard drive, or the guy writing the wiki made the same confusion as explained on the debian mailing list topic I referred to or else the Load_Cycle_Count it reports is false. It's now at 7053234 which means, if extrapolated to 24h, it would report as many load cycles per day as dario's WD Caviar during almost 2 months...

Another idea though if you really wish to reduce the amount of loading/unloading of a hard drive, shouldn't you also look at the root cause, i.e. the processes accessing the hard drive, maybe there are some kernel parameters or modules and some software configurations that reduce the frequency of system access to the hard drive.
 
 View user's profile Send private message  
Reply with quote Back to top
diblOffline
Post subject:   PostPosted: 13.04.2011, 14:51



Joined: 2010-09-12
Posts: 302
Location: Dayton, Ohio, USA
Status: Offline
While there does seem to be confusion in online discussions of load cycles, vs. start/stop cycles, Western Digital's specifications found here are very clear. They specify 600,000 load/unload cycles.
 
 View user's profile Send private message Visit poster's website  
Reply with quote Back to top
ShakaZOffline
Post subject:   PostPosted: 13.04.2011, 15:48



Joined: 2011-02-16
Posts: 67

Status: Offline
@dario : you may find this interesting http://community.wdc.com/t5/Desktop/Gre ... td-p/15731

@dibl : doesn't explain why my drive load cycles are so high. And I'm pretty sure the 600k cycles WD puts there must be a very conservative number.
 
 View user's profile Send private message  
Reply with quote Back to top
darioOffline
Post subject:   PostPosted: 13.04.2011, 16:09



Joined: 2010-11-27
Posts: 63

Status: Offline
I think it is only an approximate value, some disk may fail after 500000, another after 1000000 or more, and yes, it is probably a conservative value for obvious reasons, i'll read that thread this evening.
Btw, my problem is mainly with the disk in the acer extensa, the value increases by 20 every minute if i don't set the apm value at least to 200. I was trying to understand where to place the hdparm command to cover also the resume from suspend.
 
 View user's profile Send private message  
Reply with quote Back to top
Display posts from previous:     
Jump to:  
All times are GMT - 12 Hours
Post new topic   Reply to topic
View previous topic Printable version Log in to check your private messages View next topic
Powered by Zafenio