Re: HD spindown and fs cache writeback

[Heather <star@betelgeuse.starshine.org>]

> >I would like my HD to spin down and stay down for some time, if the
> >buffers in RAM are big enough. Therefore, I think I have to tell the
> >kernel not to write out the fs cache as often. Where can I do that?
> 
> I was wondering about this too, as I used to use a handy replacement 
> for the `update' daemon, called `mobile-update' (at 
> http://www.complang.tuwien.ac.at/ulrich/mobile-update/ )
> but this no longer works now that the bdflush daemon has vanished, 
> or the kernel interface has changed or some such.
> 
> Any ideas?  I would love to get this working again with a 2.2.15 or
> 2.3.x/2.4.x kernel, as it made a fair difference to the battery life
> and a large difference to the noise generated on some machines...

Ok, I've been looking at different (past) flavors of this same thread
recently trying to figure out the real root of the problem.

Hopefully my summation of the helpful suggestions with help some people:
	*  set noatime on filesystems where you don't care about it anyway
	*  make sure you're not syslogging things you don't care about
	*  use "noflushd"
	*  use "mobile-update"
	*  if there's some process named "sync" murder it and see where
	   its options are found so you can tweak them
	*  Use "hdparm" to advise your disk controller directly that it's
	   okay to sleep at more desirable intervals.

I'll add my own thought to the "syslog" flavor - if you *r e a l l y* don't
care, make them all log to vt's instead.  You can still look at 'em -unless-
things die, but they will never hit disk.

However in the past there have been at least a Thinkpad, an Armada, and a
couple of other systems with the problem.  Nobody posted that they succeeded
in their quest and are now happy :(  

Nor has there been a chorus of "hey, mine too!" or "yeah, we know, it's 
just busted, it sucks" that implies it happens to everyone, though perhaps 
not everyone is paying attention at this close a level.  In which case 4 or
5 over a few months might be a chorus.

So:  how does one confirm that one suffers this problem, and does APM support
need a kick in the pants to help stop it, or what?

-* Heather Stern * star@ lots of places *-