Am Samstag, 3. April 2004 02:30 schrieb Sylvain Joyeux:
Try mounting you partition using ext2 and use noflushd
DO NOT use noflushd. Noflushd captures all write access to the hd
which is a very bad idea. Even sync is captured. If you set
noflushd to 30 min. and have a crash after 29 min. then
You have with laptop_mode the same problem than with noflushd (you're
losing 29 minutes of data if your laptop crashes and the laptop_mode
timeout is 30 minutes).
Where the laptop_mode shines is in its mechanism to group write access
every time the HD has to spin up (for a read for instance)
On Fri, 2 Apr 2004 14:10:29 -0800
Jeff Coppock [EMAIL PROTECTED] wrote:
Lately, I've been working on making my laptop as efficient on power as
possible.
Dell Latitude C610, BIOS version A16, Testing release.
There is a lot of information out there on this subject and I've
accomplished
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Saturday 03 April 2004 00:10, Jeff Coppock wrote:
I've tried to see what was accessing files/dirs using the following
script:
I can get KDE to a point where nothing KDE-related shows up, but still
no joy.
I suspect FAM may be the culprit. Try
On Sat, 3 Apr 2004 00:25:43 +0200
Frans Pop [EMAIL PROTECTED] wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Saturday 03 April 2004 00:10, Jeff Coppock wrote:
I've tried to see what was accessing files/dirs using the following
script:
I can get KDE to a point where nothing
AFAIK, on ext3, the commit interval can't be changed. So, every time a
process write sth on disk, the journal commit occurs physically within
something like 10s.
Moreover, without noflushd, the kernel buffers are flushed every 30s (or
less, I can't remember)
Try mounting you partition using
On Sat, 3 Apr 2004, Sylvain Joyeux wrote:
AFAIK, on ext3, the commit interval can't be changed. So, every time a
process write sth on disk, the journal commit occurs physically within
something like 10s.
The laptop-mode patch makes the interval adjustable at mount time. The
-mm kernels
7 matches
Mail list logo
