Re: pmud options quirk
On 5 May, this message from Rogério Brito echoed through cyberspace:
> On May 05 2002, Michel Lanners wrote:
>> I've recently gotten a new HD for my TiBook (40 Gig :), which makes
>> relatively loud noise when moving the head.
> What is the manufacturer of your HD?
This one is a Fujitsu MHR2040AT. It's about 30% faster than the original
Toshiba MK1016GAP (10Gig). Idle nois is at least as low as the Toshiba;
but head movement is louder. In fact, I think it is the head assembly
moving to the parking position that makes the loud 'click' noise.
>> At first, it did that every five seconds, which, as you can imagine,
>> is unsupportable after a few minutes.
> Indeed. The same reaction that I had.
>> I've since tried to find what accesses the disk every five seconds, and
>> come to the combination pmud with APM emulation/asapm.
> That's good. I didn't investigate what was happening, but I
> guess that I'll do so when I have some spare time.
It's not that easy to see what happens. I tried to find
accessed/modified files with find -atime/-ctime, but that didn't show a
>> While we're at it, does anybody have other ideas to eliminate disk
>> acceses? I've already done these:
>> - switch back to ext2 from ext3 (ext3 prevents the HD from spinning
> I'm using ext3 (with noatime). I'll test it with ext2 to see
> the effects. But read below my comments on noatime.
>> - install noflushd
> Did that already.
Anybody know how noflushd's spin-down compares/interacts with what pmud
does in pwrctl?
>> - remount with noatime while on battery
> I always use noatime and I don't seem to have problems with
> that. But I do notice that the HD spins down when I'm on
> battery, even though I'm using ext3.
And it stays spun down for longer periods of time?
>> In particular, how about devfs? I'd suppose having devfs in memory
>> saves the atime/ctime/mtime changes on the /dev FS nodes (which is
>> metadata to the FS, right?)?
> Well, I do think that atime/ctime/mtime are metadata, but are
> they changed for device nodes? I thought that device nodes
> were special in respect to that (but I have never read any
> code to support my claim).
If I knew :). I'm now running devfs, without any problem so far (touch
wood...). For those who want to take the plunge: enabling devfs in the
kernel (but _not_ automounting it), and installing the devfsd .deb,
configuring /etc/default/devfsd to do the mount, did the trick for me.
> Ok, I know that I'm not helping, but I am just trying to see
> the similarities between our systems.
Whatever happened to that site that wanted to collect laptop-specific
tips&tricks for Mac hardware? Is taht alive somewhere?
Michel Lanners | " Read Philosophy. Study Art.
23, Rue Paul Henkes | Ask Questions. Make Mistakes.
L-1710 Luxembourg |
email email@example.com |
http://www.cpu.lu/~mlan | Learn Always. "
To UNSUBSCRIBE, email to firstname.lastname@example.org
with a subject of "unsubscribe". Trouble? Contact email@example.com