How to: Karmic Koala on Acer 1410 / 1810T/ 1810TZ

[Muskoka]

Hi, yes that's correct. I put in my password when I run the script and that's it.

Glen

[laramichaels1978]

Glad I could help! And thank YOU for curating this thread. : )

~lara

[laramichaels1978]

To build on the earlier discussion of dirty_* values, I am thinking that perhaps what really matters for avoiding spurious disk access when on battery might be dirty_background_ratio and dirty_writeback_centisecs. The former controls when (in terms of how much of total memory is dirty pages) the background kernel threads that flush stuff to disk kick in, while the latter controls how often those same threads wake up. I am guessing these two should take care of the hard drive being brought back to life every N seconds when there is "nothing much going on" (from the perspective of the user).

Of course, I know nothing about this stuff. I just read these three pages:

http://www.westnet.com/~gsmith/conte...ux-pdflush.htm

http://communities.vmware.com/thread...3B4239521977B6

http://www.gossamer-threads.com/list...kernel/1017204

dirty_ratio, on the other hand, does not seem to me to be something that triggers disk activity if the computer is largely idle; instead, it seems to be a "hard" limit that kicks in when there is so much writing going on that the background flushing mechanisms are not coping with it. When those situations happen and the system hits the value of dirty_ratio in terms of how much memory is "dirty pages", then the kernel forces that process to start syncing/flushing to disk everything it writes. This seems to be something that would happen in high-intensity scenarios, rather than a reason for the hard drive to periodically spin up while the system is mostly idle.

again, these are just uninformed guesses based on reading the above.

~l

[PatrickVogeli]

Hi Patrick,

Getting an error when I try to run the new script, after installing the other one first:

sudo ./InstallAcer_11.6_PowerSaving.sh --install_all

I then enter my password, then I get this:

################################################## ################

This script comes with no warranty. Proceed under your own
responsability and risk. If unsure, abort.
It has been tested under an Acer 1810TZ, but should work with
a wide range of intel chipset machines

################################################## ################

Enter your root password:
Password OK, proceed:

Unrecognised option, aborting...

patrick.voegeli @ gmail dot com
PatrickVogeli at ubuntuforums.org and forum.notebookreview.com

It doesn't ask for my password, this screen just comes up. What might the "Unrecognized option, aborting..." be?

Thanks for any help....Glen

Acer 1810t

./InstallAcer_11.6_PowerSaving.sh --help

as for not asking the password, it's because you are running it directly as sudo: it already has sudo privileges, so, why sould it ask for the password?. You can launch it as a regular user and it will then ask you the sudo password.

[laramichaels1978]

Finally : )

Based on the westnet.com page, I would guess that we might be msising one important parameter to avoid that constant background flushing to disk:

The first thing pdflush works on is writing pages that have been dirty for longer than it deems acceptable. This is controlled by:
/proc/sys/vm/dirty_expire_centiseconds (default 3000): In hundredths of a second, how long data can be in the page cache before it's considered expired and must be written at the next opportunity. Note that this default is very long: a full 30 seconds. That means that under normal circumstances, unless you write enough to trigger the other pdflush method, Linux won't actually commit anything you write until 30 seconds later.

So, it seems that increasing this value will be a good way to keep the flushing to disk from happening every couple of seconds.

~l

[PatrickVogeli]

Hey there,

you are good!

I'll look into it later and see if there are some scripts that already use that, so that I can get some reference values. Maybe we could set that to 30000 (5 mins)? Or should we perhaps set it at 60000 as the Dirty writeback centisecs parameter?

Your opinion is very welcome!

I'll look into this again in 12 hours.. I must go working now

[Muskoka]

Hello Patrick,

I wasn't really worried/concerned with the password. I asked in my original post, why does the script quit with "Unrecognised option, aborting..."?

Thanks Glen

[PatrickVogeli]

./InstallAcer_11.6_PowerSaving.sh --help

Look at my post.. you probably read the post and didn't download the script, then a bit later, you downloaded the script and didn't read the post again. Am I wrong?

./InstallAcer_11.6_PowerSaving.sh --install
./InstallAcer_11.6_PowerSaving.sh --install all
./InstallAcer_11.6_PowerSaving.sh --install acerhdf
./InstallAcer_11.6_PowerSaving.sh --install scripts

Run with --help to get an explanation about every option. Basically, you want to run with --install or --install all (install everything, both are the same)

[PatrickVogeli]

Finally : )

Based on the westnet.com page, I would guess that we might be msising one important parameter to avoid that constant background flushing to disk:



So, it seems that increasing this value will be a good way to keep the flushing to disk from happening every couple of seconds.

~l

after reading a bit.. I've set it to 60000 (ten minutes), so that the commit, the dirty expire and the dirty writeback all are set to ten minutes.

Thanks again

[miegiel]

after reading a bit.. I've set it to 60000 (ten minutes), so that all the commint, the dirty expire and the dirty writeback are set to ten minutes.

Thanks again

I've got the 3810T so you might want to verify it for yourself, but when I unplug the power dirty_expire_centisecs drop from 3000 to 60000 and dirty_writeback_centisecs drop from 500 to 60000 automatically.