On Wed, Dec 18, 2002 at 10:19:08AM -0500, Gary Thorpe wrote:
> [...]
> 
> I think the reason it is less severe in 1.5 (which has *again* been
> mentioned as it was in the first discussion, although the significance
> was denied then and will probably be again) is that 1.5 uses a FIXED
> size, DEDICATED file/buffer cache: applications never compete with it
> for memory, so lots of file I/O will not cause them to page out (it
> prints out the size at boot time, although 1.6 and current kernels also
> seem do this).

As has already been mentioned you can get the same behavior by proper tuning
of the VM parameters. And as has already been said, you can reproduce
this problem with 1.5.x as well, with a large enouth buffer cache, and
it's not because applications are swapped out.

Applications sometimes wants to access the disk, even if they've not been paged
out. Large X applications (like web browsers) seem to need this quite often.
When the I/O queue for a disk has a large backlog, a single access will
be delayed for a long time, blocking the application doing the access.
One side effect of UBC is that the I/O queue backlog can be much bigger than
1.5.x with the default cache size.

> 
> I/O scheduling will help I/O bound processes definately, but X or
> Mozilla are not in this category. Heavy disk activity should have

Mozilla access its file cache quite often when browsing.

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
     NetBSD: 23 ans d'experience feront toujours la difference
--