Load Cycle Count...

John Karahalis · 2009-03-03 23:00:09

Does anybody remember the load cycle bug that Ubuntu used to have? The load cycle count of laptop hard drives would increase very quickly, shortening the life of the hard drive. Slashdot has more info: http://hardware.slashdot.org/article.pl … 58&tid=198

Ubuntu ended up fixing the bug, but it seems Arch still suffers from it. The load cycle count on my new Arch laptop is increasing very quickly. I have read some discussions about this on the forum, but it didn't seem anybody came to anything conclusive.

I really just want to understand this issue once and for all so I can stop worrying about it. Is this load cycle bug a problem, and if so what can we do about it?

Dheart · 2009-03-04 05:39:15

It's not exaclty a bug, therefore there's no fixing it, rather there is hacking it. Windows machines also suffer from it. Solution for me, at least is running every time when my machine boots, or putting in /etc/rc.local this:

hdparm -B 254 /dev/sda

John Karahalis · 2009-03-04 07:42:10

Dheart wrote:

It's not exaclty a bug, therefore there's no fixing it, rather there is hacking it. Windows machines also suffer from it. Solution for me, at least is running every time when my machine boots, or putting in /etc/rc.local this:
hdparm -B 254 /dev/sda

Thanks for the help. I tried doing what you suggest back when I used Ubuntu (except using 255). Over time, any tasks related to reading/writing the hard drive really slowed the computer down. Updating my packages sometimes brought the system to a halt. I can't be 100% sure that this behavior was related, but I am fairly certain it was.

What do you mean it is not exactly a bug? Is it something I can safely ignore?

Thanks again for the response.

thunderogg · 2009-03-04 08:51:57

Look in the wiki: http://wiki.archlinux.org/index.php/Lap … wn_problem
They call it a spin-down problem, but it's not. It's just head unloads and loads. Anyway, your solution is there!!

congyiwu · 2009-03-04 09:05:26

It's not a "bug" in arch b/c it is an issue with the hard drive's firmware. A laptop hard drive when set to the default APM setting (controlled by hdparm -B) might park its heads after 20s of inactivity. If linux tends to write to /var/log/* every 30s, then the heads can park/unpark every 30s. If you had a service in windows that wrote to the disk at fixed intervals greater than the disks max idle period, you would have the same issue.

A better solution than setting hdparm -B after linux boots is to use the hard drive manufacturer's dos tools to change the APM setting to 254. I did it with my Hitachi drive. That way the setting will stick no matter what os/distro/livecd you are playing with.

John Karahalis · 2009-03-05 04:11:12

Thanks everybody for all the information.

I did try setting hdparm -B on Ubuntu. Over time, any disk-intensive work would really slow everything down. Like I said, sometimes my system would slow to a halt when doing an upgrade or working in Eclipse.

Has anybody else heard of this side-effect?

Dheart · 2009-03-05 05:01:46

I don't have this side effect. In fact, I don't notice any other difference besides the lack of clicking sound. I'm perfectly sure the problem on your ubuntu box isn't hdparm -B related, because setting power management time doesn't effect in any way the disk's performance (Of course it's "improving" performance because there is no need to "wait" for the drive to spin up anymore, but that's hardly noticeble.)

madalu · 2009-06-14 19:50:10

Sorry to resurrect an old thread, but I've just discovered this old thread with great interest. I've run smartctl -a on my three month old Asus 1000he and have discovered somewhat shocking figures:

9 Power_On_Hours 0x0032 100 100 000 Old_age Always - 555
193 Load_Cycle_Count 0x0032 077 077 000 Old_age Always - 47416

Which amounts to 85 cycles an hour!

The problem is that if I use hdaparm -B 254, I get hard drive temperatures in low to mid 40s as opposed to mid to upper 30s with a setting of 128.

I'm no expert on hard drives, so I have a simple question: Which is the greater threat to the life of the hard drive: high heat or too many load cycles?

Thanks.

Last edited by madalu (2009-06-14 19:50:57)

na12 · 2009-06-14 21:34:01

 9 Power_On_Hours          0x0012   093   093   000    Old_age   Always       -       3120
193 Load_Cycle_Count        0x0012   092   092   000    Old_age   Always       -       87103

>28, and I am not use anything

Arch Linux

#1 2009-03-03 23:00:09

Load Cycle Count...

#2 2009-03-04 05:39:15

Re: Load Cycle Count...

#3 2009-03-04 07:42:10

Re: Load Cycle Count...

#4 2009-03-04 08:51:57

Re: Load Cycle Count...

#5 2009-03-04 09:05:26

Re: Load Cycle Count...

#6 2009-03-05 04:11:12

Re: Load Cycle Count...

#7 2009-03-05 05:01:46

Re: Load Cycle Count...

#8 2009-06-14 19:50:10

Re: Load Cycle Count...

#9 2009-06-14 21:34:01

Re: Load Cycle Count...

Board footer