Nathan Scott wrote:
On Wed, Sep 08, 2004 at 03:39:55PM +0200, Jakob Oestergaard wrote:
This mail was originally sent to LKML, but it was suggested to me that I
post it here as well. Original post follows.
Heh, I should probably read all my mail before replying to LKML. :)
Thanks for forwarding.
I wonder what the effect of /proc/sys/vm/swappiness
and /proc/sys/vm/vfs_cache_pressure is on these situations.
Recent messages on the kernel channel (see thread started by
Ray Bryant from sgi), seem to suggest that the behavior
of these tunables is very variable. It should be possible
to stop the inode cache from taking over the world though.