[ubuntu_demon]

Originally Posted by Caprica_Resistance Hi, I want just to add some facts in case someone could benefit from it, because all this Load_Cycle_Count grew to some kind of urban legend or I should say "Cyber Legend". I am not professional, just user but I gathered some information: 1. No one seems to know exactly what it is all about this Load_Cycle_Count.

Getting a rough understanding about this Load_Cycle_Count issue isn't very hard. Please read the start post and all the links in it (you don't have to read about the ugly fix). Ask again if you don't understand and please be specific so I can improve this start post :
http://ubuntuforums.org/showpost.php...62&postcount=1

Originally Posted by Caprica_Resistance 2. Load_Cycle_Count is rising only on battery - wrong. On my laptop, my wife laptop, my friend laptop it is rising steadily on AC power.

On default in Ubuntu (and all other distributions as far as I know) your Load_Cycle_Count will rise with the same speed while running on AC as while running on battery. (AFAIK most operating systems won't override the defaults set by the harddrive firmware. Some BIOSes might override the default settings set by the harddrive firmware)

Having your Load_Cycle_Count rise while on battery is important because it protects the disk from bumps and saves some power. Having your Load_Cycle_Count rise while on AC is less important assuming you need less protection from bumps and you care less about power usage and care more about harddrive lifetime. Most of the time laptops are using AC so if your Load_Cycle_Count rises too fast you should probably want to reduce this speed only when on AC.

Originally Posted by Caprica_Resistance 3. Only Ubuntu is affected - wrong. On my Vista (got dual boot) Load_Cycle_Count is growing pretty fast (55.000 in 6 month), On my wife Laptop (only windows XP) Load_Cycle_Count is growing "nicely" in 1 year 77.000 - 90% of working time is AC power, on my friend laptop (mainly XP sometimes Fedora in some years above 300.000). 4. Other OS handle Load_Cycle_Count better than Ubuntu - wrong as you can see above Load_Cycle_Count can increase on various systems whether on battery or not, look at my Load_Cycle_Count changing in Fedora 8: [root@mothership daniel]# while true > do > /usr/sbin/smartctl -a /dev/sda | grep Load_Cycle_Count > sleep 300 > done 193 Load_Cycle_Count 0x0032 095 095 000 Old_age Always - 56554 193 Load_Cycle_Count 0x0032 095 095 000 Old_age Always - 56566 193 Load_Cycle_Count 0x0032 095 095 000 Old_age Always - 56579 193 Load_Cycle_Count 0x0032 095 095 000 Old_age Always - 56595 193 Load_Cycle_Count 0x0032 095 095 000 Old_age Always - 56597 193 Load_Cycle_Count 0x0032 095 095 000 Old_age Always - 56600 193 Load_Cycle_Count 0x0032 095 095 000 Old_age Always - 56607 193 Load_Cycle_Count 0x0032 095 095 000 Old_age Always - 56610 193 Load_Cycle_Count 0x0032 095 095 000 Old_age Always - 56616 193 Load_Cycle_Count 0x0032 095 095 000 Old_age Always - 56624 193 Load_Cycle_Count 0x0032 095 095 000 Old_age Always - 56635 You can see how it is increasing in every 5 minutes. The Fedora development team knows about it, but they are ignoring this fact (I read it on fedora site, can't find it right now) since (like they say) the hard disk manufacturers set this APM setting.

The amount of harddrive's power management is operating system independent.

Disk usage is depended on a lot of things. It depends on the operating system, the hardware, the applications which are run and the person using the laptop. Most Linux distributions seem to cause a bit more disk usage for example because of ext3 journaling. Most Linux distributions default to ext3 which uses journaling which is safe for your data but causes some extra disk activity. You might unknowingly have caused some unnecessary disk activity. You might have set your email client to check for email each minute. You might have set your feed reader (accidentally) to check some specific feed each minute. Some of these things might be bugs and are reported on launchpad. See : https://bugs.launchpad.net/ubuntu/+s...5/comments/185

Originally Posted by Caprica_Resistance I applied #!/bin/sh hdparm -B 254 $HDD on /etc/acpi/resume.d/ and /etc/acpi/start.d/ as recomennded: https://wiki.ubuntu.com/DanielHahler/Bug59695 and it stopped increasing though I don't know if it has any drawback.

Don't apply any unofficial ugly fixes unless you understand what you are doing and you understand how to revert. Parking the head of a laptop harddrive also protects it from bumps so you probably shouldn't turn off head parking entirely while on battery. After applying this fix keep an eye on your Load_Cycle_Count and on your harddrive temperate and make sure remains below the maximum temperature specification of your harddrive. Everything you do is on your own risk.


The biggest drawbacks of "apm 254 style fixes" are :
* no head parking : increased heat which might shorten the lifetime of your harddrive (at the very least make sure the temperature stays below the maximum temperature specification)
* no head parking : no protectection from bumps which might shorten the lifetime of your harddrive
* no head parking : increased power usage
* some people apply unofficial fixes without understanding what they are doing, without understanding how to revert them and without understanding what the drawbacks are.
* some people apply unofficial fixes without needing them (you only need them if you understand what you are doing and you will reach your Load_Cycle_Count's lifetime within three years of harddisk usage)

The advantages of using "apm 254 style fixes" :
* no head parking : a little bit better harddisk performance (is only nicer than the drawbacks if you don't move your laptop at all)
* no head parking : Load_Cycle_Count doesn't increase (is only nicer than the drawbacks if you are very heavily affected by the Load_Cycle_Count issue)

Don't bump your laptop while using apm 254. Keep an eye on your Load_Cycle_Count. Keep an eye on your temperature. Don't use apm 254 while on battery.

First read the start post and all the links in it :
http://ubuntuforums.org/showpost.php...62&postcount=1

If you would apply this fix instead you can keep your harddisk parking while on battery :
http://ubuntuforums.org/showpost.php...0&postcount=26

Originally Posted by Caprica_Resistance To MS Windows users the only method I discovered to stop Load_Cycle_Count to go higher is to play some mp3 all the time from hard disk...

Causing permanent disk activity works in all operating systems but causing permanent disk activity is bad for performance and might also decrease harddisk lifetime a little bit. Maybe the music player you use in Ubuntu uses more caching and therefor accesses the disk less.

Originally Posted by Caprica_Resistance Of course I wish next release of Ubuntu will solve this problem somehow...

Let's try to get the priority of the bug higher than wishlist.

Currently the reasoning goes that manufacturers should set appropriate settings because they should know what their hardware is capable of. But they probably don't test their harddrives with (recent) Linux Distributions so sometimes these defaults might be a bit too aggressive.

Getting a rough understanding about this Load_Cycle_Count issue isn't very hard but overriding power management defaults and reducing disk usage isn't trivial. Because it's hard to find sane defaults which would work for everybody :

The solution to the problem is hard because :
* you don't want to reduce performance too much
* you want to protect your data as much as possible (turning off journaling is not an option)
* you don't to generate too much heat (which also reduces the lifetime of your disk) and you especially don't want to cause the temperature to go above the maximum temperature specification of the disk
* you don't to use too much power
* you want to protect the harddisk from bumps (especially when on battery)
* finding sane apm defaults is hard

Finding sane apm defaults is very hard because :
* 128 is standardized as safing as much power as possible without spinning down (unless you set the spin down time) so parking the head as much as possible. But how much head parking is different for each disk.
* 254 is standardized as using no power management and no head parking at all (which doesn't protect your harddrive from bumps and might generate too much heat). * Values in between 128 and 254 are different for each disk.